Connect with us

Published

on

Meta has built custom computer chips to help with its artificial intelligence and video-processing tasks, and is talking about them in public for the first time.

The social networking giant disclosed its internal silicon chip projects for the first time to reporters earlier this week, ahead of a Thursday virtual event discussing its AI technical infrastructure investments.

related investing news

Nvidia's stock could rise fivefold in 10 years on A.I. trend, says fund manager

CNBC Pro

Investors have been closely watching Meta’s investments into AI and related data center hardware as the company embarks on a “year of efficiency” that includes at least 21,000 layoffs and major cost cutting.

Although it’s expensive for a company to design and build its own computer chips, vice president of infrastructure Alexis Bjorlin told CNBC that Meta believes that the improved performance will justify the investment. The company has also been overhauling its data center designs to focus more on energy-efficient techniques, like liquid cooling, to reduce excess heat.

One of the new computer chips, the Meta Scalable Video Processor (MSVP), is used to process and transmit video to users while cutting down on energy requirements. Bjorlin said “there was nothing commercially available” that could handle the task of processing and delivering 4 billion videos a day as efficiently as Meta wanted.

The other processor is the first in the company’s Meta Training and Inference Accelerator (MTIA) family of chips intended to help with various AI-specific tasks. The new MTIA chip specifically handles “inference,” which is when an already-trained AI model makes a prediction or takes an action.

Bjorlin said that the new AI inference chip helps power some of Meta’s recommendation algorithms used to show content and ads in people’s news feeds. She declined to answer who is manufacturing the chip, but a blog post said that the processor is “fabricated in TSMC 7nm process,” indicating that chip-giant Taiwan Semiconductor Manufacturing is producing the technology.

She said that Meta has a “multi-generational roadmap” for its family of AI chips that include processors used for the task of training AI models, but declined to offer details beyond the new inference chip. Reuters previously reported that Meta cancelled one AI inference chip project and started another that was supposed to roll out around 2025, but Bjorlin declined to comment on that report.

Because Meta isn’t in the business of selling cloud computing services like companies including Google-parent Alphabet or Microsoft, the company didn’t feel compelled to publicly talk about its internal data center chip projects, she said.

“If you look at we’re sharing—our first two chips that we developed—it’s definitely giving a little bit of a view into what are we doing internally,” Bjorlin said. “We haven’t had to advertise this, and we don’t need to advertise this, but you know, the world is interested.”

Meta vice president of engineering Aparna Ramani said the company’s new hardware was developed to work effectively with its home-grown PyTorch software, which has become one of the most popular tools used by third-party developers to create AI apps.

The new hardware will eventually be used to power tasks related to the metaverse, such as virtual reality and augmented reality, as well as the burgeoning field of generative AI, which generally refers to AI software that can create, compelling text, images, and videos.

Ramani also said that Meta has developed a generative AI-powered coding assistant for the company’s developers to help them more easily create and operate software. The new assistant is similar to Microsoft’s GitHub Copilot tool that it released in 2021 with help from the AI startup OpenAI.

In addition, Meta said it completed the second-phase buildout, or the final buildout, of its supercomputer dubbed Research SuperCluster (RSC), which the company detailed last year. Meta used the supercomputer, which contains 16,000 Nvidia A100 GPUs, to train the company’s LLaMA language model, among other uses.

Ramani said that Meta continues to act on its belief that it should contribute to open-source technologies and AI research in order to push the field of technology. The company has disclosed that its biggest LLaMA language model, LLaMA 65B, contains 65 billion parameters and was trained on 1.4 trillion tokens, which refers to the data used for AI training.

Companies like OpenAI and Google have not publicly disclosed similar metrics for their competing large language models, although CNBC reported this week that Google’s PaLM 2 model was trained on 3.6 trillion tokens and contains 340 billion parameters.

Unlike other tech companies, Meta released its LLaMA language model to researchers so they can learn from the technology. However, the LlaMA language model was then leaked to the wider public, leading to many developers building apps incorporating the technology.

Ramani said that Meta is “still thinking through all of our open source collaborations, and certainly, I want to reiterate that our philosophy is still open science and cross collaboration.”

Watch: A.I. is a big driver of sentiment for big tech

Continue Reading

Technology

Tesla must pay portion of $329 million in damages after fatal Autopilot crash, jury says

Published

on

By

Tesla must pay portion of 9 million in damages after fatal Autopilot crash, jury says

A jury in Miami has determined that Tesla should be held partly liable for a fatal 2019 Autopilot crash, and must compensate the family of the deceased and an injured survivor a portion of $329 million in damages.

Tesla’s payout is based on $129 million in compensatory damages, and $200 million in punitive damages against the company.

The jury determined Tesla should be held 33% responsible for the fatal crash. That means the automaker would be responsible for about $42.5 million in compensatory damages. In cases like these, punitive damages are typically capped at three times compensatory damages.

The plaintiffs’ attorneys told CNBC on Friday that because punitive damages were only assessed against Tesla, they expect the automaker to pay the full $200 million, bringing total payments to around $242.5 million.

Tesla said it plans to appeal the decision.

Attorneys for the plaintiffs had asked the jury to award damages based on $345 million in total damages. The trial in the Southern District of Florida started on July 14.

The suit centered around who shouldered the blame for the deadly crash in Key Largo, Florida. A Tesla owner named George McGee was driving his Model S electric sedan while using the company’s Enhanced Autopilot, a partially automated driving system.

While driving, McGee dropped his mobile phone that he was using and scrambled to pick it up. He said during the trial that he believed Enhanced Autopilot would brake if an obstacle was in the way. His Model S accelerated through an intersection at just over 60 miles per hour, hitting a nearby empty parked car and its owners, who were standing on the other side of their vehicle.

Naibel Benavides, who was 22, died on the scene from injuries sustained in the crash. Her body was discovered about 75 feet away from the point of impact. Her boyfriend, Dillon Angulo, survived but suffered multiple broken bones, a traumatic brain injury and psychological effects.

“Tesla designed Autopilot only for controlled access highways yet deliberately chose not to restrict drivers from using it elsewhere, alongside Elon Musk telling the world Autopilot drove better than humans,” Brett Schreiber, counsel for the plaintiffs, said in an e-mailed statement on Friday. “Tesla’s lies turned our roads into test tracks for their fundamentally flawed technology, putting everyday Americans like Naibel Benavides and Dillon Angulo in harm’s way.”

Following the verdict, the plaintiffs’ families hugged each other and their lawyers, and Angulo was “visibly emotional” as he embraced his mother, according to NBC.

Here is Tesla’s response to CNBC:

“Today’s verdict is wrong and only works to set back automotive safety and jeopardize Tesla’s and the entire industry’s efforts to develop and implement life-saving technology. We plan to appeal given the substantial errors of law and irregularities at trial.

Even though this jury found that the driver was overwhelmingly responsible for this tragic accident in 2019, the evidence has always shown that this driver was solely at fault because he was speeding, with his foot on the accelerator – which overrode Autopilot – as he rummaged for his dropped phone without his eyes on the road. To be clear, no car in 2019, and none today, would have prevented this crash.

This was never about Autopilot; it was a fiction concocted by plaintiffs’ lawyers blaming the car when the driver – from day one – admitted and accepted responsibility.”

The verdict comes as Musk, Tesla’s CEO, is trying to persuade investors that his company can pivot into a leader in autonomous vehicles, and that its self-driving systems are safe enough to operate fleets of robotaxis on public roads in the U.S.

Tesla shares dipped 1.8% on Friday and are now down 25% for the year, the biggest drop among tech’s megacap companies.

The verdict could set a precedent for Autopilot-related suits against Tesla. About a dozen active cases are underway focused on similar claims involving incidents where Autopilot or Tesla’s FSD— Full Self-Driving (Supervised) — had been in use just before a fatal or injurious crash.

The National Highway Traffic Safety Administration initiated a probe in 2021 into possible safety defects in Tesla’s Autopilot systems. During the course of that investigation, Tesla made changes, including a number of over-the-air software updates.

The agency then opened a second probe, which is ongoing, evaluating whether Tesla’s “recall remedy” to resolve issues with the behavior of its Autopilot, especially around stationary first responder vehicles, had been effective.

The NHTSA has also warned Tesla that its social media posts may mislead drivers into thinking its cars are capable of functioning as robotaxis, even though owners manuals say the cars require hands-on steering and a driver attentive to steering and braking at all times.

A site that tracks Tesla-involved collisions, TeslaDeaths.com, has reported at least 58 deaths resulting from incidents where Tesla drivers had Autopilot engaged just before impact.

Read the jury’s verdict below.

Continue Reading

Technology

Crypto wobbles into August as Trump’s new tariffs trigger risk-off sentiment

Published

on

By

Crypto wobbles into August as Trump's new tariffs trigger risk-off sentiment

A screen showing the price of various cryptocurrencies against the US dollar displayed at a Crypto Panda cryptocurrency store in Hong Kong, China, on Monday, Feb. 3, 2025. 

Lam Yik | Bloomberg | Getty Images

The crypto market slid Friday after President Donald Trump unveiled his modified “reciprocal” tariffs on dozens of countries.

The price of bitcoin showed relative strength, hovering at the flat line while ether, XRP and Binance Coin fell 2% each. Overnight, bitcoin dropped to a low of $114,110.73.

The descent triggered a wave of long liquidations, which forces traders to sell their assets at market price to settle their debts, pushing prices lower. Bitcoin saw $172 million in liquidations across centralized exchanges in the past 24 hours, according to CoinGlass, and ether saw $210 million.

Crypto-linked stocks suffered deeper losses. Coinbase led the way, down 15% following its disappointing second-quarter earnings report. Circle fell 4%, Galaxy Digital lost 2%, and ether treasury company Bitmine Immersion was down 8%. Bitcoin proxy MicroStrategy was down by 5%.

Stock Chart IconStock chart icon

hide content

Bitcoin falls below $115,000

The stock moves came amid a new wave of risk off sentiment after President Trump issued new tariffs ranging between 10% and 41%, triggering worries about increasing inflation and the Federal Reserve’s ability to cut interest rates. In periods of broad based derisking, crypto tends to get hit as investors pull out of the most speculative and volatile assets. Technical resilience and institutional demand for bitcoin and ether are helping support their prices.

“After running red hot in July, this is a healthy strategic cooldown. Markets aren’t reacting to a crisis, they’re responding to the lack of one,” said Ben Kurland, CEO at crypto research platform DYOR. “With no new macro catalyst on the horizon, capital is rotating out of speculative assets and into safer ground … it’s a calculated pause.”

Crypto is coming off a winning month but could soon hit the brakes amid the new macro uncertainty, and in a month usually characterized by lower trading volumes and increased volatility. Bitcoin gained 8% in July, according to Coin Metrics, while ether surged more than 49%.

Ether ETFs saw more than $5 billion in inflows in July alone (with just a single day of outflows of $1.8 million on July 2), bringing it’s total cumulative inflows to $9.64 to date. Bitcoin ETFs saw $114 million in outflows in the final trading session of July, bringing its monthly inflows to about $6 billion out of a cumulative $55 billion.

Don’t miss these cryptocurrency insights from CNBC Pro:

Continue Reading

Technology

Google has dropped more than 50 DEI-related organizations from its funding list

Published

on

By

Google has dropped more than 50 DEI-related organizations from its funding list

Google CEO Sundar Pichai gestures to the crowd during Google’s annual I/O developers conference in Mountain View, California, on May 20, 2025.

David Paul Morris | Bloomberg | Getty Images

Google has purged more than 50 organizations related to diversity, equity and inclusion, or DEI, from a list of organizations that the tech company provides funding to, according to a new report.

The company has removed a total of 214 groups from its funding list while adding 101, according to a new report from tech watchdog organization The Tech Transparency Project. The watchdog group cites the most recent public list of organizations that receive the most substantial contributions from Google’s U.S. Government Affairs and Public Policy team.

The largest category of purged groups were DEI-related, with a total of 58 groups removed from Google’s funding list, TTP found. The dropped groups had mission statements that included the words “diversity, “equity,” “inclusion,” or “race,” “activism,” and “women.” Those are also terms the Trump administration officials have reportedly told federal agencies to limit or avoid.

In response to the report, Google spokesperson José Castañeda told CNBC that the list reflects contributions made in 2024 and that it does not reflect all contributions made by other teams within the company.

“We contribute to hundreds of groups from across the political spectrum that advocate for pro-innovation policies, and those groups change from year to year based on where our contributions will have the most impact,” Castañeda said in an email.

Organizations that were removed from Google’s list include the African American Community Service Agency, which seeks to “empower all Black and historically excluded communities”; the Latino Leadership Alliance, which is dedicated to “race equity affecting the Latino community”; and Enroot, which creates out-of-school experiences for immigrant kids. 

The organization funding purge is the latest to come as Google began backtracking some of its commitments to DEI over the last couple of years. That pull back came due to cost cutting to prioritize investments into artificial intelligence technology as well as the changing political and legal landscape amid increasing national anti-DEI policies.

Over the past decade, Silicon Valley and other industries used DEI programs to root out bias in hiring, promote fairness in the workplace and advance the careers of women and people of color — demographics that have historically been overlooked in the workplace.

However, the U.S. Supreme Court’s 2023 decision to end affirmative action at colleges led to additional backlash against DEI programs in conservative circles.

President Donald Trump signed an executive order upon taking office in January to end the government’s DEI programs and directed federal agencies to combat what the administration considers “illegal” private-sector DEI mandates, policies and programs. Shortly after, Google’s Chief People Officer Fiona Cicconi told employees that the company would end DEI-related hiring “aspirational goals” due to new federal requirements and Google’s categorization as a federal contractor.

Despite DEI becoming such a divisive term, many companies are continuing the work but using different language or rolling the efforts under less-charged terminology, like “learning” or “hiring.”

Even Google CEO Sundar Pichai maintained the importance diversity plays in its workforce at an all-hands meeting in March.

“We’re a global company, we have users around the world, and we think the best way to serve them well is by having a workforce that represents that diversity,” Pichai said at the time.

One of the groups dropped from Google’s contributions list is the National Network to End Domestic Violence, which provides training, assistance, and public awareness campaigns on the issue of violence against women, the TTP report found. The group had been on Google’s list of funded organizations for at least nine years and continues to name the company as one of its corporate partners.

Google said it still gave $75,000 to the National Network to End Domestic Violence in 2024 but did not say why the group was removed from the public contributions list.

WATCH: Alphabet’s valuation remains highly attractive, says Evercore ISI’s Mark Mahaney

Alphabet's valuation remains highly attractive, says Evercore ISI's Mark Mahaney

Continue Reading

Trending