Generative AI has been a hot topic, especially after the launch of ChatGPT by OpenAI. It has even exceeded Metaverse in popularity. From top tech firms like Google, Microsoft and Adobe to chipmakers like Qualcomm, Intel, and NVIDIA, all are integrating generative AI models in their products and services. So, why is generative AI attracting interest from all these companies?
While generative AI and ChatGPT are both used for generating content, what are the key differences between them? The content generated can include solutions to problems, essays, email or resume templates, or a short summary of a big report to name a few. But it also poses certain challenges like training complexity, bias, deep fakes, intellectual property rights, and so on.
In the latest episode of ‘The Counterpoint Podcast’, host Maurice Klaehne is joined by Counterpoint Associate Director Mohit Agrawal and Senior Analyst Akshara Bassi to talk about generative AI. The discussion covers topics including the ecosystem, companies that are active in the generative AI space, challenges, infrastructure, and hardware. It also focuses on emerging opportunities and how the ecosystem could evolve going forward.
Tesla delivered nearly 343,900 vehicles during Q3 2022, an increase of 42.4% YoY
Logistics remains a major bottleneck for Tesla deliveries
Tesla can exceed 1.3 million unit deliveries by year end with current trajectory
Tesla rebounded during Q3, after experiencing a relatively weak second quarter. During Q3, Tesla delivered nearly 343,900 vehicles, a 42.4% annual increase and a sequential increase of 35%. The combined deliveries of Model S and Model X grew by more than 100% YoY, reaching 18,670 units, while the combined deliveries of Model 3 and Model Y increased by 40% YoY. China is the leading market for Tesla followed by the USA and Europe.
Tesla’s Shanghai Gigafactory surpassed the previous quarterly production rate and remains the main export hub supplying to most markets outside North America. The gigafactory updated its production ramp in July this year. The Berlin Gigafactory is also producing more than 2,000 units of Model Y, weekly. A lot of work is left to bring the Berlin plant to full capacity as it is only slowly reaching its planned output. As winter approaches, and it is feared that Europe will experience an energy crisis, Musk somehow remains optimistic about vehicle production in the Berlin plant and expects that no production cuts will happen.
Q3 financial summary:
During Q3, Tesla’s total revenue grew by almost 56% YoY, reaching $21.4 billion. Tesla generated $18.6 billion from the vehicle segment, an increase of 55% YoY. This is largely due to increased global deliveries and higher vehicle ASPs.
Although revenue from vehicle leasing during Q3 has increased significantly by 61% YoY, revenue from the sale of automotive credits grew by just 2.5% YoY.
Revenue generated from the company’s other businesses like energy storage, solar panel deployment, charging and vehicle servicing also grew by 62.5% YoY, exceeding $2.7 billion.
Gross profit, was $5.3 billion an increase of 47% YoY. But below expectation due to the high cost of raw materials, upgrading the production ramps (Berlin, Texas and 4680 cell factories) and increased logistic costs.
Tesla has been facing a serious issue with vehicle deliveries. There weren’t enough transport vehicles available with its logistic partners to handle the volume of Tesla deliveries. This increases the logistic cost which, in turn, is affecting the per-vehicle cost.
3.4% of the total revenue has been diverted towards R&D expenditure during Q3 2022. R&D spending stood at $0.73 billion, an increase of 20% YoY and sequentially growth of 10%. This is apparently due to the development of Tesla’s Optimus Robot and full-self driving (FSD) capability. This year Tesla postponed its AI Day to showcase a working prototype of its humanoid Optimus Robot whose software is very similar to the FSD system.
The FSD beta users reached 160,000 in Q3, up from 100,000 in Q2. Tesla is also going for a wider release of its FSD beta during Q4 2022. Hence, new Tesla owners will have the option to avail FSD beta immediately. Currently, there is an eligibility criteria to avail the FSD beta. With the resignation of Andrej Karpathy, Tesla’s AI and Autopilot director, it was perceived that the company’s FSD development is likely to stall, but it seems Tesla has made good progress and is confident of its path toward full autonomy, despite some alarming failures among beta testers.
Outlook:
Despite a weak second quarter, Tesla’s yearly deliveries may cross 1.3 million units by the end of 2022. Tesla is expected to make its first delivery of the Tesla Semi truck to Pepsi on December 1st this year. The Semi is claimed to have a range of 500 miles with cargo at ground level. We are also expecting to see the company’s long-advertised Cybertrucks becoming available by mid-2023. Alongside these, Tesla has also increased the production of its in-house designed 4680 cells. The constant production ramp upgrade in its gigafactories around the globe is likely to keep Tesla the market leader in the battery electric vehicle segment.
Google introduced its affordable Pixel “a” series smartphone in 2019, focusing on offering pure Android OS, Google smarts, and flagship computational photography experiences at lower price points. In 2020, we used the Pixel 4a (review) and were impressed with its still photography capabilities. Now, we have the latest $450 Google Pixel 6a which brings several improvements over the predecessor.
The Pixel flagship smartphones up to the Pixel 5 were powered by Qualcomm’s Snapdragon 8-series SoC, whereas the “a” series devices were powered by a mid-or-high tier Snapdragon 6/7-series SoC. But that changes this year. In late 2021, Google introduced the Pixel 6 & 6 Pro smartphones powered by Google’s own custom silicon, the Tensor SoC. The Pixel 6a is also powered by the same flagship Tensor SoC, offering smart AI & ML capabilities at a lower price point. Below is our analysis of the Google Pixel 6a after using it for over a month.
WATCH: Google Pixel 6a Review
Compact form factor, refreshing dual-tone finish
The Pixel 6a has a more compact form factor than some competitor smartphones in the same price segment. This is thanks to the 6.1-inch OLED screen with a 60Hz refresh rate, a 20:9 aspect ratio, and 1080×2400 pixels resolution. Even though the refresh rate is just 60Hz, the software and hardware optimizations offer a fluid experience.
The smartphone has an aluminum alloy frame with curved edges offering a comfortable fit when you hold the device. Corning Gorilla Glass 3 protection is upfront, whereas the back is made from a transparent polycarbonate sheet that looks and feels like glass.
Just like the previous Pixel smartphones, the 6a also has a two-tone finish on the back, giving it a distinct look. The dual camera modules and LED flash sit slightly raised in the strip offering it a neat look. Google has also added an IP67 rating for water resistance, meaning it can survive under one-meter water for up to 30 minutes. And it is a good addition to a smartphone in this segment.
Google Tensor: Snappy Android performance, lots of smart AI features
One of the biggest highlights of the Pixel 6a is its Tensor SoC. Built on Samsung’s 5nm process node, it is a joint effort between Google and Samsung to enable machine learning and advanced computational photography experiences.
The CPU of this custom silicon offers a 2+2+4 core configuration, which is different from the 1+3+4 cores that we have seen on competitor chips. So, you get two Cortex-X1 cores clocked at 2.8GHz to do all the heavy lifting, and two Cortex-A76 cores clocked at 2.25GHz. Then there are four Cortex-A55 efficiency cores clocked at 1.8GHz. On the graphics side, you have a 20-core Mali G78 MP20 GPU for gaming and other intensive tasks. There is also Google’s Titan M2 co-processor adding a level of privacy and security to your data.
All that power from the flagship chipset along with 6GB of RAM, offers smooth and snappy performance. We played games like Call of Duty: Mobile & Asphalt 9: Legends, both of which ran smoothly. Even after 20 minutes of gaming, the device did not get warm, thus offering good thermal management.
Even with a few apps open in the background, multi-tasking is a breeze. Scrolling through the UI, social media feed, surfing favorite websites even with a few tabs open, gaming, and much more, the Pixel 6a can handle all this without any hiccup.
The Tensor SoC offers snappy performance with good thermal management.
The Google Tensor SoC unlocks a host of new experiences: –
New videography experiences:
Pixel smartphones have been known for offering a great still camera, but with the powerful Tensor chip, the videography experience has improved too. The Cinematic Pan mode slows down the panning movements to offer professional-looking cinematic shots. We shot a few cinematic videos, and they look awesome.
The “Speech Enhancement” feature in the camera app is an impressive addition, and great for vloggers. Here’s a sample clip where you can see how the noise-cancelling works well, focusing on the speech. #Pixel6a#CPReviewspic.twitter.com/8BuLg9Reaj
There is also a new Speech Enhancement feature that uses AI & Machine Learning to reduce the ambient background noise and focus on your speech, offering crisp and clear audio. This is a great and handy feature for vloggers and content creators. We tested this feature and were left quite impressed.
Speech Enhancement is a neat feature that vloggers will appreciate.
Object removal quickly gets rid of photobombers:
How many times have you come across scenarios where you visited a great place, and clicked a nice photo, but it got ruined by a photobomber? Well, Google’s Object Removal tool is now available in the Photos app. Simply select the photo > Tap on Edit > Tools > Magic Eraser.
The #Pixel6a powered by Google’s #Tensor SoC brings some cool AI & ML features such as Object Eraser, Blur, Colour Pop & Camouflage. Here is a quick look at how it works, showcasing how far AI has come. #CPReviewspic.twitter.com/HVbLVXysSG
By default, it will offer suggestions to remove unwanted objects or people from the frame. But you can also circle the objects and the AI will do its job. The feature is not 100% accurate but gets very close. The best part of the object removal feature is that it works on older photos too. Back in 2017, I visited Universal Studios in Singapore and clicked a few photos, but all had photobombers. Thankfully, five years later, I can now get rid of them, and it works brilliantly. The photos were shot on a Samsung Galaxy Note8. Below is a sample photo showcasing Google Tensor’s AI Prowess.
There is one more tick Google has up its sleeve in the Photos app, it’s called the camouflage feature. It is available in the Magic Eraser tool. According to Google, the feature lets you “change the color of an object so it pops out or blends into its surroundings.” It works pretty well, but on all photos.
The Magic Eraser feature offers a good demonstration of AI & ML capabilities.
Other than these, you also get other features like Blur & Colour-pop that have been around for a while, but with Tensor, the processing gets a little faster. With “Blur” feature, you can convert any photo into a portrait mode photo, by adding a nice background blur. The setting is available in the Photos app under Tools, and you can also control the depth and intensity of the blur. Lastly, there is also “Color Pop” feature that lets you keep the subject in color, and the rest of the background is in black & white. Both features work well, and below are some samples.
Built-in Live Translate:
The Pixel 6a comes with a Live Translate feature within the messaging app. This allows you to speak in your preferred language, say English, and translate it into any desired language. We tried speaking in English and translating it into Hindi, Marathi, Chinese and Spanish language, and it seems to work well. Thanks to Tensor SoC, the translation speed and accuracy are also improved.
In the messaging app, you can also tap the mic button and dictate the message. So, you don’t have to go through the hassle of typing long messages. With commands like Delete, Clear and Clear All, you can delete the last word, clear a sentence, or clear all text as well.
Besides, the Tensor SoC is also better at handling “ambient experiences” such as the “Now Playing” feature without draining the battery.
Live Translate & Speech-to-Text feature is very handy and could be helpful when traveling, or when talking to people with language barriers.
Great still cameras, good videography; but…
Google has once again succeeded in delivering a fantastic still camera, while also improving the videography experience. Daylight shots from both wide and ultrawide cameras have ample details, and the dynamic range is good too.
It is impressive how Google has been using the same primary camera since the Pixel 2 and still manages to bring improvements every year. The camera lets you capture portrait shots and 2X digital zoom photos with good details too. Below are some camera samples.
Ultrawide
Wide
2X Digital Zoom
The wide and ultrawide cameras capture great details, both in daylight and low-light conditions.
Both the ultrawide and wide cameras capture good low-light photos too. With Night Sight mode, the cameras capture more than the naked eye can see. Below are some samples.
While both cameras are great for capturing landscapes, the AI sometimes gets too aggressive and adds shadows to people’s skin tone. This is prominently seen when capturing photos against the sunlight. With HDR, the AI tries to get most elements right; like clouds, flowers, a glass of wine, and so on; but in that attempt, the skin tone gets a few shades darker. Fine-tuning the algorithms could possibly fix these issues. Below is an example showing how AI gets too aggressive.
The Pixel 6a offers great still photography capabilities, but the algorithms need fine-tuning to deliver consistent results.
In proper lighting conditions, the front camera also does a good job of clicking detailed selfies. The skin tones are close to natural, details are good and portrait mode also offers good separation of background and foreground. samples photo below.
In terms of videography, there are dedicated modes like Speech Enhancement & Cinematic Pan which we spoke about above. But the general videography experience is also much improved. Videos at FHD (1080p) resolution are very well stabilized, both at 30fps and 60fps. At 4K resolution too, both at 30fps and 60fps, the video is decently stabilized. However, frame drops and jerks are noticeable when you are shooting while running. But overall, you get a better videography experience at this price point compared to some competitors.
OS & battery life: stock Android never felt better
Coming to the software, the Pixel 6a ships with Android 12 out-of-the-box. But the Android 13 update is already available to download and installs right after booting the device. Android 12 was a major departure from previous Android versions, adding more customizations and enhanced privacy features. Android 13 brings further enhancements such as auto theming icons for third-party apps, more cosmetic additions to the Material You theming, and a double line clock on the lock screen to name a few.
But that’s not all, there are many under-the-hood changes, including bug fixes and performance improvements. Bluetooth LE and Spatial Audio support, Privacy, and Security updates offering more user controls are some of the other features of Android 13. The media player now has a cleaner layout, the clipboard now comes with nice visualizations. Also, you now get to edit the copied text in the clipboard. Lastly, you can also assign a specific language to different apps, which could be a handy feature for users in countries like India.
Google is promising three major Android OS upgrades and five years of security updates for the Pixel 6a.
In terms of battery life, the Pixel 3a and Pixel 4a would barely make it to the end of the workday with moderate-to-heavy usage. However, the Tensor SoC brings big efficiency gains, and even with heavy usage, you will still be left with enough battery before going to bed. Our Pixel 6a offers a screen time of just over five hours with moderate to heavy usage, which seems good enough for a phone with a 4,410mAh battery.
With the Tensor SoC, the Pixel 6a sees good improvements in battery life compared to the previous Pixel “a” series smartphone.
Similar to most OEMs today, Google is also not bundling a charging adapter, but you do get a Type-C to Type-C data cable in the box. The smartphone supports up to 18W fast charging speeds, and with some 60W fast chargers lying around, it took nearly one hour to charge the smartphone from almost empty to full.
Key Takeaways
• The Pixel 6a’s two-tone back finish and the camera strip offer it an attractive look and feel.
• The Tensor SoC adds value by bringing smart AI & ML experiences.
• The SoC also offers snappy performance, security, and improved battery life.
• The Pixel 6a offers great photography and videography experiences, however, Google needs to fine-tune the algorithms to improve skin tones.
• Theming and privacy features of Android 13 add a refreshing feel and safety to the stock Android experience.
• The Pixel 6a supports 5G (Sub-6GHz) and comes with eSIM capabilities, which is good to see in this segment.
• The optical fingerprint scanner is slow and sometimes fails in reading the fingerprint.
ALSO READ: Other Strategic Reviews on Smartphones, Smartwatches, TWS & More
Tesla sold more than 254,000 vehicles in Q2 2022, an increase of 27% YoY, which was below general expectations.
This was the first time since the COVID-19-hit 2020 that the automaker experienced a sequential decline in sales.
After achieving phenomenal growth in Q1 2022, Tesla’s global sales during Q2 2022 grew by just 27% YoY to over 254,000 units, falling short of expectations. In QoQ terms, the sales fell 18%. Business during Q2 2022 was affected by COIVD-19-related shutdowns in China. Production units in and around Shanghai were closed temporarily due to strict lockdown measures. As a result, Tesla sold just 89,000 cars across China during Q2 2022. Cumulative sales in China during April and May fell by more than 66% YoY. The situation improved only after the production returned to full capacity in June.
It was expected that the Berlin Gigafactory would boost Tesla’s sales in Europe after becoming operational in March 2022. But the production was lower than expected. A few rumored reasons for the low production are litigation with the German government and a shortage of human resources. The Berlin factory is currently focusing on the production and deliveries of the Model Y across Europe.
Tesla bets on in-house battery cell manufacturing
Tesla delivered its first batch of cars equipped with the in-house 4680 battery cells and structural battery packs during this quarter. These cells use a little amount of lithium. With lithium prices soaring worldwide, 4680 cells will help lower the vehicle manufacturing cost. The cells will power the Model Ys coming out of the Berlin Gigafactory. However, Tesla will shut the Berlin Gigafactory for a couple of weeks during autumn to upgrade the production system of 4680 cells.
Other businesses see 33% YoY growth
Although Tesla’s vehicle sales in Q2 2022 failed to meet expectations, its other businesses like energy deployment and storage, charging and other services grew more than its vehicle segment. Energy deployment, energy storage, charging and other services grew by 33% YoY. Tesla deployed 106 MW of solar panels and 1.13 GWh of energy storage during Q2 2022. It installed 247 new superchargers worldwide, bringing its global supercharger number to 3,971 units with more than 36,000 connectors.
Tesla converts 75% of its Bitcoins to fiat currency
During Q2 2022, Tesla also converted 75% of its Bitcoins to fiat currency. This was done to have a better cash position against the backdrop of COIVD-19-related uncertainties. This conversion reduced Tesla’s digital assets to $218 million and added $936 million in cash to Tesla’s balance sheet.
Source: Tesla Q2 2022 Financials and Counterpoint Analysis
Q2 2022 Financial Results
During Q2 2022, Tesla sold more than 254,000 vehicles at 27% YoY growth. The Model 3 and Model Y comprised more than 93% of these sales.
Revenue from vehicle sales stood at $14.6 billion. Total revenue grew by almost 42% YoY, with the COVID-19 impact on China reducing the QoQ number by about 10%. Revenue generated from automotive credit also declined slightly compared to Q2 2021.
The company’s other services, like energy storage, charging and insurance, contributed to 14% of its total revenue. Revenue from insurance and vehicle services saw a 54.2% YoY growth, while the energy storage and charging segment grew by just 8% YoY. The energy storage business was expected to perform better but was restricted due to semiconductor-related supply issues.
Tesla’s gross profit during Q2 2022 reached $4.2 billion and stood at 25%. Though the shutdown in China adversely affected the business, increase in US deliveries along with the higher average vehicle price helped Tesla earn 47% more profit YoY.
R&D costs grew 16% YoY during Q2 2022. Tesla is trying to achieve complete autonomy by 2024 by perfecting Full-Self Driving (FSD) software. But the resignation of Andrej Karpathy, the director of artificial intelligence and autopilot system at Tesla, in mid-July is likely to stall the progress of this project, which is expected to get delayed by a year.
Source: Tesla Q2 2022 Financials and Counterpoint Analysis
Market Outlook
Despite experiencing a dip during the second quarter of 2022, Tesla’s future outlook seems strong and promising with strong fundamentals. Tesla has secured the supply of LFP batteries for its Shanghai Gigafactory by signing a deal with BYD. Transitioning to LFP batteries and 4680 battery cells will help Tesla reduce vehicle manufacturing costs. Moreover, Tesla expects the Berlin Gigafactory production capacity to cross 100,000 units by the end of 2022. With all these developments, Tesla is expected to cross more than 1.2 million units of vehicle deliveries by the end of 2022.
The use of IoT devices is making our daily lives smarter. IoT combined with artificial intelligence (AI), or artificial IoT (AIoT), is helping in automated real-time decision-making and data analysis. AI can add value to IoT through machine learning (ML) and improved decision-making. Similarly, IoT can add value to AI through connectivity and data exchange. With the rapid technological advancements, AIoT is transforming every industry, enterprise and consumer.
5G is going to be the key ingredient in driving AIoT applications. According to a Counterpoint Research study, the shipments of 5G AIoT-supported modules will grow at a CAGR of 84% between 2022 and 2030. Further, 60% of 5G IoT modules will have AI capability by 2030 for better processing and real-time decision-making.
AIoT applications
The adoption of AIoT is becoming an emerging technology trend across a wide range of industries where real-time data operation is needed, such as industrial manufacturing, robotics, logistics, healthcare, agriculture, smart cities and smart home.
Industrial manufacturing
Manufacturing relies on digital transformation to become more efficient and reduce human error. This sector needs to adopt AIoT solutions. AIoT-powered robots in factories improve the manufacturing process with excellent efficiency. They help reduce labor costs as well as time.
Smart cities
In a smart city, there are several uses of AIoT, such as traffic management and waste management. To avoid chaos and congestion on the road in a crowded city, AIoT-based drones help monitor traffic and transmit real-time traffic data for analyses through AI and for making decisions on the speed limit and timing of traffic lights, all without human interference. Therefore, real-time traffic monitoring by drones increases efficiency and reduces congestion.
Security and surveillance are other important applications for AI in smart cities. AI cameras can help police monitor illegal activities and prevent unwanted situations.
Autonomous vehicles
Self-driving cars are one of the best use cases of AIoT applications. IoT-enabled devices like cameras, radars and sonars in the car gather data and the AI system helps analyze this data within a few milliseconds so that the car can make decisions like a human. Fully autonomous vehicles will generate 1-2TB of data per hour and AI will be required to handle this amount of data and take some decisions at the edge.
Smart homes
AI in the connected home space is mainly used for voice assistant, situational awareness, automation and security. Starting from door locks, smoke alarm, surveillance and smart speakers to smart appliances such as lighting, thermostat, refrigerator, plugs, routers, meters, home controllers and vehicle chargers, many applications have already adopted AI features.
Challenges in implementing AI in IoT
As most AI applications are based on real-time decision-making, they need a high-speed data rate to communicate. 5G’s high speed and low latency will be ideal for AIoT applications. But in many regions, 5G infrastructure is still not there or is in the initial phase. Hence, it will be challenging to scale. Data management and taking the right decision at the right time by handling huge amounts of data will be another big challenge for AI adoption. Moreover, the security angle will also need to be addressed. Both hardware- and software-level security will be required for AIoT applications to keep connected devices safe.
Initiatives by module and chipset players in AI applications
With the increasing traction for 5G-based AIoT applications, module vendors like Fibocom, Quectel, Thundercomm and MeiG are stepping forward to launch AI-supported 5G IoT modules. Some module vendors are offering AI features at the hardware level while some vendors are offering AI features at the software level. Whether to use hardware- or software-level AI features depends on the application and cost of the project. Recently, Quectel announced that it would add software-level AI capability in its Rel 16-based 5G modules by partnering with NVIDIA. International module vendor Telit is also adding AI capabilities in its FM980 5G module through NVIDIA software. However, Quectel, Fibocom, MeiG and Thundercomm are already offering hardware-level 5G AI-supported modules for high-end applications such as C-V2X, AR/VR, robots, smart cities, live streaming, gaming and edge computing.
In terms of chipset players, Qualcomm is leading in the 5G AIoT chipset space. This year, it launched the world’s first AI-supported 5G modem Snapdragon X70. Moreover, it has broader 5G AIoT SoC offerings with the QCM6490, Snapdragon 480, Snapdragon 690 and Snapdragon 750.
The second-largest IoT chipset player, UNISOC, is trying to gain momentum for 5G through AIoT-based SoCs. So far, UNISOC is offering AIoT features in its T770, T760 and T740 chipsets.
Recently, MediaTek launched the Genio 1200 chipset, specially designed for 5G AIoT devices. It is targeting applications such as smart home, industrial, robotics and audio/video terminals.
We expect that 2023 will provide momentum to the 5G AIoT market as the IoT market has been facing some instability lately due to inflation, supply issues and other macro factors.
The Pixel 6 and Pixel 6 Pro smartphones come with a new and impressive industrial design.
The smartphones are powered by Google’s in-house Tensor SoC that leverages AI and ML to enable rich new experiences.
After officially announcing its Tensor processor in August this year to power the Pixel smartphones, Google has now launched the Pixel 6 and Pixel 6 Pro models. Marking a departure from the mid-range hardware in last year’s Pixel 5, the new Pixel 6 series brings top-notch hardware while focusing on performance, photography and AI smarts. With the Pixel 6 starting at $599 and the Pixel 6 Pro starting at $899, what improvements do they bring over the Pixel 5? Let us deep-dive to see what Google has on offer with the new Pixel smartphones.
Pixel 6, Pixel 6 Pro: Stunning Design, Upgraded Camera Sensors
Since the first Pixel, Google has always done a terrific job when it comes to industrial design. The same continues with the Pixel 6 series featuring a thick, symmetrical camera bar that puts the camera front and center. The dual-tone design language offers a sophisticated look and feel, making the Pixel 6 smartphones stand out from the crowd.
Source – Google
The lower half of the back below the camera bar has a lighter color shade, whereas the upper half has a darker color shade. These color accents make the Pixel 6 and Pixel 6 Pro look stunning.
Moving to displays, the Pixel 6 comes with a 6.4-inch FHD+ AMOLED screen with 90Hz refresh rate. The Pixel 6 Pro, on the other hand, comes with a bigger 6.7-inch QHD+ display with 120Hz refresh rate. Both panels offer a variable refresh rate that can automatically adjust based on the on-screen content.
Source – Google
In terms of configuration, the Pixel 6 has 8GB of RAM (LPDDR5) and two on-board storage options to choose from – 128GB and 256GB (UFS 3.1). The Pixel 6 Pro has 12GB of RAM and three storage options to choose from – 128GB, 256GB and 512GB. It is good to see Google keeping with the current memory trends and offering more on-board storage, especially as it no longer offers unlimited Google Photos storage.
Moving to photography, the new Pixel 6 series gets big camera upgrades. The Pixel 6 comes with a dual-camera setup, featuring a 50MP wide primary camera having f/1.9 aperture, laser auto-focus and OIS. Using a quad bayer filter that combines four adjacent pixels into one, the camera outputs 12.5MP images. The Pixel 6 also comes with a secondary ultrawide camera of 12MP resolution and field of view of 114 degrees, which is a little less compared to the 120 degrees in the iPhone 13 series, and 123 degrees in the Samsung Galaxy Z Fold 3.
Source – Google
The Pixel 6 Pro comes with a triple rear camera setup, where it shares the primary wide and secondary ultrawide sensors with the Pixel 6. The third is a 48MP f/2.5 aperture folded telephoto lens for periscope style zoom. Having a focal length of 104mm, the camera allows for up to 4x optical zoom, and 20x digital zoom with Google’s ‘Super Res Zoom’ feature. Upfront, the Pixel 6 comes with an 8MP selfie camera, whereas the 6 Pro model comes with an 11MP front shooter. There are a bunch of camera features that the Tensor processor unlocks, which we will talk about in a bit.
Besides still photography, Google has also focused on the videography aspect with the Pixel 6 devices. Both smartphones can record 4K 60fps videos from the rear cameras, and 240fps slow-motion videos, along with 4K timelapse and cinematic pan. The Pixel 6 front camera can record 1080p videos in up to 60fps, whereas the Pixel 6 Pro can record 1080p 60fps and also 4K 30fps videos.
Source – Google
Rest of the specifications include dual SIM support with one physical SIM and one eSIM. Connectivity-wise, both Pixel 6 smartphones support 5G (Sub-6GHz and mmWave), Wi-Fi 6E, Bluetooth 5.2, wireless charging and IP68 certification for water and dust resistance. The Pixel 6 has a 4,610mAh battery, whereas the Pixel 6 Pro comes with a slightly larger 5,000mAh battery. Both support 30W fast charging and 20W wireless fast charging. Both Pixels ship with Android 12 out-of-the-box.
In terms of pricing, there is almost a $300 price delta between the Pixel 6 at $599 and the Pixel 6 Pro at $899. Although the pricing is still aggressive compared to the other latest Samsung and OnePlus flagship devices, limited differentiation between the two variants makes the Pixel 6 Pro a hard sell.
In the US, the Pixel 6 variant will compete in the affordable premium category, which saw a huge uptick in demand last year driven by the Galaxy S20 FE. US carriers are already offering good trade-in value and other promotional deals on the purchase of a new Pixel 6 device. In such a scenario, the Pixel 6 can gather strong momentum ahead of the holiday season, given it is able to maintain a strong supply at the time when the industry is facing severe shortages,” said Hanish Bhatia, senior analyst at Counterpoint Research.
Google Tensor Processor: Unlocking Smart AI and Computational Photography Experiences
Besides the new industrial design and upgraded camera hardware, the custom-built Tensor SoC is one of the big talking points of the new Pixels. Building its own custom processor speaks a lot about Google’s ambition towards becoming a vertically integrated player like Apple or Samsung, and it makes sense too. This allows Google to develop its own processor baked with impressive AI capabilities to differentiate itself from the competition. Another benefit is that the device and chipset launch timelines can be in sync, unlike earlier where the Snapdragon-powered Pixel would get outdated within a month with Qualcomm launching a new 800-series chipset in December.
“Designing own chipset gives full control to smartphone makers like Samsung and Google to add their own customizations while reducing dependence on chipmakers. Through Tensor, Google aims to have full control of the entire stack from chip to OS to middleware to cloud. Depending on the success of the in-house designing of the Tensor SoC, Google can also scale it to other products such as Google Assistant-powered smart speakers and Nest devices for edge inferencing,” said Tarun Pathak, research director at Counterpoint.
Talking about the Tensor SoC in detail, Google has made some interesting choices. Made on Samsung’s 5nm process, the Tensor SoC uses ARM’s Cortex CPU cores in tri-cluster 2+2+4 configuration:
Two Cortex-X1-based performance cores clocked at 2.8GHz.
Two Cortex-A76-based medium cores clocked at 2.2GHz.
Four Cortex-A55-based efficiency cores clocked at 1.8GHz.
20-Core ARM Mali-G78 MP20 GPU
Google says that the Tensor SoC offers up to 80% faster CPU performance and 370% faster GPU performance over the Pixel 5.
Source – Google
The Google Tensor SoC also comes with a Tensor Processing Unit (TPU) for Machine Learning (ML) tasks, and an advanced image signal processing unit (ISP) for computational photography improvements. There is a Context Hub that enables low-power activities such as always-on display, now playing andassistant, and other ambient experiences. Alongside the Tensor Security core is a Titan M2 co-processor. Google says the Pixel 6 devices have most layers of hardware security.
“While there will be some apprehensions around the new Google Tensor chip, Pixel users will be able to get firsthand experience of Google’s AI prowess powered by its on-board TPU. Further enhancements in computational photography with more accurate skin tones, reduced image blur and new features such as Magic Eraser and Motion Mode, are expected to remain key selling points for the device,” added Hanish Bhatia
Google highlighted some of the breakthrough AI and ML features it is bringing to the Pixel 6 devices with the new chipset. In photography, it is bringing features like Magic Eraser, Face Unblur, Motion Mode and Real Tone.
How many times have your perfect photos been ruined by photobombers? After clicking a photo, the Magic Eraser tool will let you remove unwanted people/objects and fill in that place to look like they never existed. This is not a new feature. We have seen a similar implementation in Nokia Lumia devices before, and even in the likes of the Samsung Galaxy S4 (Drama Shot). While that was years ago, the latest implementation is likely to reflect improvements in AI and imaging since then.
Source – Google
There are times when we click photos of children where they are not able to stand still, resulting in blurry faces. The Pixel 6 takes two images simultaneously from the main and ultra-wide cameras. The ultrawide camera uses a faster exposure to reduce blur. The ML then fuses the sharper face from the ultrawide camera, thus offering a blur-free photo. Next is Motion Mode, which takes several photos, then determines the subject, figures out what is moving and adds static blur to the background. All this without needing fancy equipment to capture long exposure shots.
But one of the most interesting features is Real Tone. The AI systems on most smartphone cameras are designed keeping lighter complexion people in mind, or to make people with a dark complexion look brighter. Google, using ML and computational photography,improves exposure and white balance to better represent people of every skin tone.
Source – Google
With the Tensor SoC, Google is also making typing much faster on the Pixel 6 with the speech-to-text mode. From understanding the accent and dialect to isolating the voice, a lot of things play an important role in making natural language processing more robust and accurate. With the Tensor SoC, Google says voice typing is much faster and effortless.
Now, when you call a bank’s customer care, you have to wait till you connect to the executive, hear the menu options carefully and so on. Google aims to improve that experience by showing the expected wait time before you dial the number. Also, the IVR menu options will be transcribed and shown on the screen, so you can simply tap and select the right option.
Google has also included a Live Translation feature that works in the messaging app. If you want to reply to someone in Japanese, you can simply speak in English and the app will then translate it to Japanese. You can then hit the send button. This does away with the need to open a translation app, copy-paste the text in the message box and so on. This translation feature is also available in WhatsApp, Google Chat, Instagram, Twitter and other apps.
Lastly, there is also an interpreter mode that enables a live translation feature, helpful when taking interviews or traveling. All these features are a great improvement on paper, and it will be interesting to see how these work in real life.
Key Takeaways:
The Pixel 6 series is a big upgrade in terms of design and hardware compared to previous Pixel devices.
Google is betting heavily on new capabilities offered by the Tensor SoC as a way to differentiate its offerings from the competition.
The starting price of $599 for the Pixel 6 makes it an attractive option for those who want to try out the best of Google’s hardware and software improvements.
As home security solutions such as security cameras are now used in several households, there is an increasing need for AI (Artificial Intelligence) inferencing at the edge. Typically, the security solution relies on machine learning models to identify objects or faces in the CCTV footage – for example differentiating between a cat and a human. The data is then sent to the cloud for analysis and sent back to the device. But this can be problematic when there is poor internet connectivity.
Ergo, a tiny 7x7mm Edge AI chip from a company called Perceive, aims to solve issues with AI inferencing at the edge. The chip enables rapid processing on edge devices, for example facial recognition, or alerting to certain sounds, such as glass breaking or a dog barking. This can trigger actions without resorting to cloud-based systems. This type of solution can also offer enhanced data security and user privacy, as the data does not leave the device. The edge AI inference chips can be used in connected devices such as smart speakers as well, where many commands can be processed on the device, rather referring to the cloud. There can be many other applications in the future including drones, autonomous vehicles, and much more.
In the latest episode of ‘The Counterpoint Podcast’, host Peter Richardson is joined by David McIntyre, VP of Marketing at Perceive. David talks about AI inferencing at the edge using a tiny chip called Ergo. He deep dives into problems solved by inferencing on edge devices over the cloud, use cases, and savings made related to space onboard, costs and power. The podcast discussion also focuses on potential applications where solutions like Perceive’s Ergo chip can be used.
If you have not heard about audio-based social media platforms like Clubhouse, Twitter Spaces and Facebook Live Audio, you are probably living under a rock. Humans, being social animals, need conversations. But the ongoing pandemic has made it a bit cumbersome to have them, just the right time for audio-based social media platforms to make an entry and take the internet by storm.
If the pandemic wasn’t enough to drive this growth, Tesla chief Elon Musk’s tweets about going live on Clubhouse added warp speed to it. According to Clubhouse CEO Paul Davidson, the app has 10 million weekly active users currently. Unofficial estimates put the startup’s value at over $4 billion.
Social Media Giants Late to the Party?
It is surprising how the incumbent social media giants missed this bus. After all, they spend a lot of time and money to identify the next growth opportunity to drive user engagement, which ultimately translates into ad revenue. But Clubhouse was not the first missed opportunity for them. Earlier, Instagram came up with stories which we now see across all social media platforms, including Twitter and LinkedIn. TikTok also disrupted social media with short-format videos and Instagram was quick to follow. But while large tech corporations can get blindsided in identifying elements of their growth strategy, their reach and scale give them the privilege to be late to the market and still remain a leader.
Can Voice be a Revenue Engine for Social Media, Streaming and Messaging Platforms?
Audio-based social media has drawn interest outside of social media too. Apart from Facebook Live Audio and Twitter Spaces, music-streaming giant Spotify has launched Greenroom, gaming chat app Discord has launched Stage, and Telegram has tweaked its audio features for live voice chat. Other players like Racket, Fireside, Soapbox and Air Time are also attempting to capture value from more customized experiences.
While everyone competes for monthly active users (MAUs) and user engagement, there are several opportunities to monetize the ever-growing audio-based social media.
Paid and brand sponsored community rooms could open a revenue opportunity for audio-based social media platforms. YouTube already has a community of content creators who monetize video content. Instagram influencers leverage their reach for sponsored brand posts while Discord channels have done the same with information and file sharing. Voice is no different and seems highly promising to become the next growth avenue for content creators. How about Spotify being a platform for live singing?
Similarly, brand sponsored community rooms can foster new ideas for marketers. They can connect with customers and open new interactions that could lead to stronger brand loyalty. Brands can further leverage voice platforms for product launch, flash sales, brand trivia contests, promotions, and much more.
Additionally, millions of daily conversations could be a goldmine. Voice data could be used to reveal and index user interests much more precisely than other user data. Synthesis of voice conversations has the potential to unlock billions in ad revenue.
Audio-based social media thrives on the intimacy, originality and immediacy that voice interactions provide. At present, audio-based social media is at a very early stage and has a long journey ahead. It will continue to evolve and create several monetization opportunities across the social media and other connectivity platforms. But who wins the race will solely depend on scalability, time to market and product innovation.
One in three smartphones to be shipped in 2020 will natively embed machine learning and artificial intelligence (AI) capabilities at the chipset level. Apple, with its Bionic system on chip (SoC), proliferating across its complete portfolio over the next couple of years, will drive native AI adoption in smartphones. Its universal adoption of AI-capable SoCs will likely enable Apple to lead the AI-capable chip market through 2020.
Huawei, with its HiSilicon Kirin 970 SoC, launched in September and finding application in the Huawei Mate 10 series launched today in Munich, is second to market after Apple with AI-capable smartphones. The Huawei Mate 10 is able to accomplish diverse computational tasks efficiently, thanks to the neural processing unit at the heart of the Kirin 970 SoC.
However Qualcomm will unlock AI capabilities in its high to mid-tier SoCs within the next few months. It should be able to catch-up and is expected to be second in the market in terms of volume by 2020, followed by Samsung and Huawei.
Exhibit 1: Native AI capable smartphone shipment forecast by 2020
Machine learning and AI have not made major headway in smartphone applications until the second half of 2017 due to the limited processing power of smartphone CPUs, meaning the user experience would have been hindered. AI applications require huge amounts of data processing even for a small task. Sending and receiving that information from cloud-based data centers is potentially difficult, time consuming and requires a solid connection, which is not always available. The answer is to have the AI-capability on-board the device.
The initial driver for the rapid adoption of AI in smartphones is the use of facial recognition technology by Apple in its recently launched iPhone X. Face recognition is computationally intensive and if other vendors are to follow Apple’s lead, they will need to have similar on-board AI capabilities to enable a smooth user experience.
With advanced SoC-level AI capabilities, smartphones will be able to perform a variety of tasks such as processing natural languages, including real-time translation; helping users take better photos by intelligently identifying objects and adjusting camera settings accordingly. But this is just the start. Machine learning will make smartphones understand user behaviour in an unprecedented manner. Analysing user behaviour patterns, devices will be able to make decisions and perform tasks that will reduce physical interaction time between the user and the device. Virtual assistants will become smarter by analysing and learning user behaviour, thereby uniquely serving each user according to their needs. This could potentially help virtual assistants take the leap and become a main-stream medium of interaction between the user and device.
Native AI can also effectively counter the increasing security threat smartphones are facing by things like, real-time malware detection, recognizing user behaviour to identify if the phone is being misused, analysing email and other apps for things like phishing attacks.There is also growing potential in for AI-capable devices to play a key role in health care. Machine learning algorithms can be used to generate health and lifestyle guidance for users by analysing combinations of sensor data and user behaviour.
Overall, we expect AI-capable smartphones to proliferate rapidly at the top-end of the market, but to relatively quickly filter into mid-range devices from the mid to latter part of 2018. By 2020 we expect over one third of all smartphones shipping to be natively AI-capable.
Over Half a Billion Smartphones Will Be Shipped with Out-of-the-Box On-Board AI Capabilities
According to the latest research from Counterpoint’s Components Tracker Service, one in three smartphones to be shipped in 2020 will natively embed machine learning and artificial intelligence (AI) capabilities at the chipset level. Apple, with its Bionic system on chip (SoC), proliferating across its complete portfolio over the next couple of years, will drive native AI adoption in smartphones. Its universal adoption of AI-capable SoCs will likely enable Apple to lead the AI-capable chip market through 2020.
Huawei, with its HiSilicon Kirin 970 SoC, launched in September and finding application in the Huawei Mate 10 series launched today in Munich, is second to market after Apple with AI-capable smartphones. The Huawei Mate 10 is able to accomplish diverse computational tasks efficiently, thanks to the neural processing unit at the heart of the Kirin 970 SoC.
However Qualcomm will unlock AI capabilities in its high to mid-tier SoCs within the next few months. It should be able to catch-up and is expected to be second in the market in terms of volume by 2020, followed by Samsung and Huawei.
Exhibit 1: Native AI capable smartphone shipment forecast by 2020
Machine learning and AI have not made major headway in smartphone applications until the second half of 2017 due to the limited processing power of smartphone CPUs, meaning the user experience would have been hindered. AI applications require huge amounts of data processing even for a small task. Sending and receiving that information from cloud-based data centers is potentially difficult, time consuming and requires a solid connection, which is not always available. The answer is to have the AI-capability on-board the device.
Commenting on the analysis, Counterpoint Research Director, Jeff Fieldhack noted, ‘The initial driver for the rapid adoption of AI in smartphones is the use of facial recognition technology by Apple in its recently launched iPhone X. Face recognition is computationally intensive and if other vendors are to follow Apple’s lead, they will need to have similar on-board AI capabilities to enable a smooth user experience.’
Adding to this, Research Director, Peter Richardson, said ‘With advanced SoC-level AI capabilities, smartphones will be able to perform a variety of tasks such as processing natural languages, including real-time translation; helping users take better photos by intelligently identifying objects and adjusting camera settings accordingly. But this is just the start. Machine learning will make smartphones understand user behaviour in an unprecedented manner. Analysing user behaviour patterns, devices will be able to make decisions and perform tasks that will reduce physical interaction time between the user and the device. Virtual assistants will become smarter by analysing and learning user behaviour, thereby uniquely serving each user according to their needs. This could potentially help virtual assistants take the leap and become a main-stream medium of interaction between the user and device.’
Research Analyst, Shobhit Srivastava, commented further, ‘Native AI can also effectively counter the increasing security threat smartphones are facing by things like, real-time malware detection, recognizing user behaviour to identify if the phone is being misused, analysing email and other apps for things like phishing attacks.’
‘There is also growing potential in for AI-capable devices to play a key role in health care. Machine learning algorithms can be used to generate health and lifestyle guidance for users by analysing combinations of sensor data and user behaviour.’
‘Overall, we expect AI-capable smartphones to proliferate rapidly at the top-end of the market, but to relatively quickly filter into mid-range devices from the mid to latter part of 2018. By 2020 we expect over one third of all smartphones shipping to be natively AI-capable.’
Background:
Counterpoint Technology Market Research is a global research firm specializing in Technology products in the TMT industry. It services major technology firms and financial firms with a mix of monthly reports, customized projects and detailed analysis of the mobile and technology markets. Its key analysts are experts in the industry with an average tenure of 13 years in the high-tech industry.
In order to access
Counterpoint Technology Market Research Limited (Company or We hereafter) Web sites, you may be asked to complete a registration form. You are required to provide contact information which is used to enhance the user experience and determine whether you are a paid subscriber or not.
Personal Information
When you register on we ask you for personal information. We use this information to provide you with the best advice and highest-quality service as well as with offers that we think are relevant to you. We may also contact you regarding a Web site problem or other customer service-related issues. We do not sell, share or rent personal information about you collected on Company Web sites.
How to unsubscribe and Termination
You may request to terminate your account or unsubscribe to any email subscriptions or mailing lists at any time.
In accessing and using this Website, User agrees to comply with all applicable laws and agrees not to take any action that would compromise the security or viability of this Website. The Company may terminate User’s access to this Website at any time for any reason. The terms hereunder regarding Accuracy of Information and Third Party Rights shall survive termination.
Website Content and Copyright
This Website is the property of Counterpoint and is protected by international copyright law and conventions. We grant users the right to access and use the Website, so long as such use is for internal information purposes, and User does not alter, copy, disseminate, redistribute or republish any content or feature of this Website. User acknowledges that access to and use of this Website is subject to these TERMS OF USE and any expanded access or use must be approved in writing by the Company.
– Passwords are for user’s individual use
– Passwords may not be shared with others
– Users may not store documents in shared folders.
– Users may not redistribute documents to non-users unless otherwise stated in their contract terms.
Changes or Updates to the Website
The Company reserves the right to change, update or discontinue any aspect of this Website at any time without notice. Your continued use of the Website after any such change constitutes your agreement to these TERMS OF USE, as modified.
Accuracy of Information:
While the information contained on this Website has been obtained from sources believed to be reliable, We disclaims all warranties as to the accuracy, completeness or adequacy of such information. User assumes sole responsibility for the use it makes of this Website to achieve his/her intended results.
Third Party Links:
This Website may contain links to other third party websites, which are provided as additional resources for the convenience of Users. We do not endorse, sponsor or accept any responsibility for these third party websites, User agrees to direct any concerns relating to these third party websites to the relevant website administrator.
Cookies and Tracking
We may monitor how you use our Web sites. It is used solely for purposes of enabling us to provide you with a personalized Web site experience.
This data may also be used in the aggregate, to identify appropriate product offerings and subscription plans. Cookies may be set in order to identify you and determine your access privileges. Cookies are simply identifiers. You have the ability to delete cookie files from your hard disk drive.
