Big Tech's $600 billion spending plans exacerbate investors' AI headache

The two artificial intelligence startups behind rival chatbots ChatGPT and Claude are bracing for an existential showdown this year as both need to prove they can grow a business that will make more money than they’re losing.

The fiercest competition between the two AI developers, along with bigger companies like Google, is a race to win over corporate leaders looking to adopt AI tools to boost workplace productivity. The rivalry is also spilling into other realms, including the Super Bowl.

Anthropic is airing a pair of TV commercials during Sunday’s game that ridicule OpenAI for the digital advertising it’s beginning to place on free and cheaper versions of ChatGPT. While Anthropic has centered its revenue model on selling Claude to other businesses, OpenAI has opened the doors to ads as a way of making money from the hundreds of millions of consumers who get ChatGPT for free.

Anthropic’s commercials humorously mock the dangers of manipulative chatbots — represented as real people speaking in a stilted and unnaturally effusive tone — that form a relationship with a user before trying to hawk a product. The commercials end with a written message — “Ads are coming to AI. But not to Claude.” — followed by the opening beat and lyrics of the Dr. Dre song “What’s the Difference."

In a sign they struck a nerve, OpenAI CEO Sam Altman said in a social media post that he laughed at the “funny” ads but blasted them as dishonest and threw shade at his competitor’s smaller customer base.

“Anthropic serves an expensive product to rich people,” Altman wrote on X. He also boasted that more Texans “use ChatGPT for free” than all the people in the United States who use Claude.

Chiming in to directly challenge Anthropic CEO Dario Amodei was OpenAI’s president and co-founder Greg Brockman, who questioned whether Anthropic was truly committing to never selling Claude “users’ attention or data to advertisers.” Amodei, who rarely posts on X, did not respond.

The rivalry has existed ever since Amodei and other OpenAI leaders quit the AI research laboratory and formed Anthropic in 2021, promising a clearer focus on the safety of the better-than-human technology called artificial general intelligence that both San Francisco firms wanted to build. That was before OpenAI first released ChatGPT in late 2022, revealing the huge commercial potential of large language models that could help write emails, homework, or computer code.

The competition ramped up this week as both companies launched product updates. OpenAI on Thursday launched a new platform called Frontier, designed to be a one-stop shop for businesses adopting a variety of AI tools, including those not made by OpenAI, that can work in tandem. It’s part of a push toward AI agents that work autonomously as “AI co-workers” on someone’s behalf.

“We can be the partner of choice for AI transformation for enterprise. The sky is the limit in terms of revenue we can generate from a platform like that,” Fidji Simo, OpenAI’s CEO of applications, told reporters this week.

Anthropic, in turn, on Thursday announced an upgrade to its “smartest model,” claiming that the new Claude Opus 4.6 “plans more carefully, sustains agentic tasks for longer, operates reliably in massive codebases, and catches its own mistakes.” OpenAI followed that shortly after with yet another update, a new version of its Codex coding tool, it says, that can “do nearly anything” professionals do on a computer.

“Both OpenAI and Anthropic are really trying to position themselves as a platform company,” said Gartner analyst Arun Chandrasekaran. “The models are important, but the models aren’t a means to an end.”

The two startups aren’t just competing with each other. They also face competition from Google, which is both a leading developer of a powerful AI model, Gemini, and has its own cloud computing infrastructure backed by revenue from its legacy digital advertising business. They also have complicated relationships with Amazon, which is Anthropic’s primary cloud provider, and Microsoft, which holds a 27% stake in OpenAI.

The first choice for businesses looking to adopt AI agents is typically cloud computing “hyperscalers” like Microsoft, Google, and Amazon, which offer a package of services, while AI model providers like Anthropic and OpenAI “tend to come in second place,” said Nancy Gohring, a senior research director at IDC.

But there’s an opening because none of the players are giving businesses what they want, which are stronger security and compliance assurances to enable the more widespread use of AI agents that can access corporate systems and data.

“Adopting AI and agents is inherently somewhat risky,” Gohring said.

There’s also the AI division of Elon Musk’s newly merged SpaceX and its chatbot, Grok, which is not yet a viable contender for business customers. Musk has long set his sights on challenging the market dominance of OpenAI, which he co-founded and is now suing in a court case set for trial in April.

SpaceX, OpenAI, and Anthropic are among the world’s most valuable privately held firms, and Wall Street investors expect any or all of them could become publicly traded within the next year or so. But unlike SpaceX, which has its rocket business to fall back on, or established tech giants — like Amazon, Googl,e and Microsoft — both Anthropic and OpenAI must find a way to make enough from selling AI products to pay for the huge costs in computer chips and data centers to run their energy-hungry AI systems.

It’s not that Anthropic and OpenAI aren’t making money or growing their product lines. The private firms don’t publicly disclose sales, but both have signaled they are making billions of dollars in revenue on their existing products, including paid chatbot subscriptions for individual users.

But it costs a lot more money to fund the computing infrastructure needed to build these powerful AI models and respond to the millions of prompts they get each day. OpenAI, in particular, has said it owes more than $1 trillion in financial obligations to backers — including Oracle, Microsoft, and Nvidia — that are essentially fronting the compute costs on the expectation of future payoffs.

For some, the wait will likely be worth it.

“Profitability matters, but not as a near‑term decision factor for investors who remain focused on scale, differentiation, and infrastructure leverage,” said Forrester analyst Charlie Dai. “Both companies continue to post heavy losses, yet investors still back them because the frontier‑model race demands extraordinary capital intensity.”

Denise Dresser, OpenAI’s newly hired chief revenue officer, told reporters this week that the company’s priority is “building the best enterprise platform for all industries, all segments.”

“I don’t think we’re thinking about it from a revenue standpoint, but truly from a customer outcome standpoint,” she said, in part reflecting the “sense of urgency” she’s heard from CEOs who want a smoother way of applying AI.

“There’s a recognition that AI is becoming a core operating advantage,” Dresser said. “They don’t want to be on the wrong side of that shift.”

If you think we’ve hit sports saturation, well, you ain’t seen nothing yet. Between the Super Bowl and the Winter Olympics, this weekend concentrates one of the largest overlapping live audiences in the world, spanning broadcast, streaming, and social platforms globally. Here’s what I’m watching for:

🏈 The Super Bowl as a platform and beyond

The Super Bowl remains the biggest stage in American advertising. What’s changed is how that stage is being used. Brands are still using the Super Bowl to broadcast a message, but they’re also demonstrating cultural fluency through the partnerships they choose and the formats they adopt. The most interesting work feels less like a traditional broadcast spot and more like brands stepping into existing formats that audiences already trust.

For example, rather than attempting to cram corporate messaging into 30 seconds, Salesforce has handed over creative control to MrBeast. In typical MrBeast fashion, he’s engineering a contest with cash prizes. The collaboration is turning a Super Bowl ad into a participatory game, a format that MrBeast has mastered. This is a smart way for Salesforce to gain more cultural currency with younger audiences through a trusted influencer.

📺 How NBC manages the Olympics and the Super Bowl

NBC is airing both the Super Bowl and the Winter Olympics, which is pretty much a media bonanza spanning linear and streaming audiences. NBC is orchestrating attention across platforms. Peacock acts as the connective tissue, carrying live Olympic events, whip-around coverage, and rapid highlight packages while linear TV anchors the Super Bowl (although audiences do overlap considerably between linear and streaming). Viewers are encouraged to move fluidly between events during commercial breaks, halftime, and natural lulls, reinforced by highlights, push notifications, and social clips that guide rather than interrupt behavior.

Brands can’t treat these events as isolated buys. A Super Bowl spot now has to assume the viewer may also be watching Olympics coverage, scrolling highlights, or jumping between feeds. The Olympics complement the Super Bowl by offering more frequent touchpoints tied to live moments, athlete storytelling, and real-time cultural relevance. I am watching to see how brands design campaigns that travel across that flow: a broadcast moment that points to a streaming extension, social content that activates during game breaks, and creative that works whether it’s seen live, clipped, or discovered hours later.

A planned $600 billion artificial intelligence spending splurge by big tech firms in 2026 is adding to investor unease as they assess the implications for profitability, as well as a potential existential threat to software firms.

Shares of Amazon (AMZN.O), opens new tab, which had announced a $200 billion capital expenditure outlay, slid 7% on Friday, while Alphabet (GOOGL.O), opens new tab lost 3% after the company said on Wednesday that capital spending could double this year. Meta Platforms (META.O), opens new tab, was down 1.3%.

Other heavyweight technology companies, however, were trading higher: Nvidia (NVDA.O), opens new tab rose 7%, Microsoft (MSFT.O), opens new tab gained 1% and Tesla (TSLA.O), opens new tab was up 4%. The benchmark S&P 500 (.SPX), opens new tab added 1.6% while the Nasdaq rose 2%, although both indexes are set to finish the week lower.

"The market's viewpoint is that the AI build-out trade, and the way they've pulled forward all these earnings for many, many years, we think that's just got too pricey," said Andrew Wells, chief investment officer at SanJac Alpha in Houston. "It's not that the trade is over, but it got too pricey in pulling forward all these potential future revenues and not really pricing in the risk into all that. So it's a de-risking trade."

NVIDIA CEO Jensen Huang attributed the uptick in spending to "sky-high" demand. Speaking on CNBC's "Halftime Report", he called the rise appropriate and sustainable.

Meanwhile, the equities of data analytics firms continued to come under selling pressure on concerns that they face an existential threat from powerful new AI models.

Canada-based Thomson Reuters (TRI.TO), opens new tab, (TRI.N), opens new tab, which suffered a record one-day plunge earlier this week, was down 0.7%. London-listed RELX's (REL.L), opens new tab shares lost 4.6% and notched a 17% tumble in their worst week since 2020.

The S&P 500 software and services index (.SPLRCIS), opens new tab has fallen almost 8% this week and has seen around $1 trillion in market value evaporate since January 28.

"Headlines that would have pushed shares to fresh highs during the peak of AI optimism are now being interpreted far more cautiously by investors," said Carlota Estragues Lopez, equity strategist at St. James's Place in London.

"It's not just return-on-investment that worries investors, but also the risk of narrow market leadership that struggles to broaden beyond a handful of mega-cap names."

Tech giants collectively expected to spend at least $630 billion this year on AI

JOLT TO DATA ANALYTICS FIRMS

A selloff in software, data, and analytics firms was triggered by a new plug-in from Anthropic's Claude.

Shares in London Stock Exchange Group (LSEG.L), opens new tab, clawed back some ground on Friday, but their price was still down almost 8% for the week in a second straight week of sharp losses.

This week's drawdown in AI-exposed shares has weighed on broader equity markets. Global shares (.MIWD00000PUS), opens new tab are on track to ease 0.33% for the week.

The rout has been particularly acute in India, where shares of software exporters plunged another 2% on Friday as they ended a week that has seen $22.5 billion in market value losses.

Investor nerves over potential AI‑driven disruption are coinciding with a growing tendency to punish big tech firms for signaling even heavier spending on the technology.

Google parent Alphabet also upped its spending plans on Thursday, sending its shares as much as 8% lower at one point, although they ended the day flat.

"Both Alphabet and Amazon delivered strong underlying business performance, driven by better-than-expected growth in cloud. But that hasn't been enough to distract markets from their ballooning capital investment plans," said Aarin Chiekrie, equity analyst, Hargreaves Lansdown.

The word on the Chinese street is that metals are the next hot thing. All you need to join the bull party is an online trading account, a bit of cash for the initial margin, and access to the right WeChat chatroom to meet like-minded punters.

China's metal exchanges have, for two months, been struggling to contain the unprecedented liquidity rush caused by the spreading metals mania.

The Shanghai Futures Exchange and the Guangzhou Futures Exchange have, between them, raised margins and tightened trading rules 38 times to maintain order.

The spate of interventions has covered the metallic spectrum from precious metals, gold and silver, to industrial inputs such as nickel and lithium.

China has a long history of manias in markets as obscure as ferro-silico,n but nothing on this scale.

Moreover, metal fever has swept up the rest of the world. The speculative stampede has generated extreme volatility in the silver market and rocked even safe-haven gold.

ShFE volumes of aluminium, copper, nickel and tin

MASS MOMENTUM

The Shanghai Futures Exchange registered record levels of trading activity across multiple metal contracts in January.

Tin volumes exceeded one million metric tons on a single day as the price accelerated to fresh all-time highs. That's more than twice the world's annual physical usage.

The crowding of retail investors acts as a giant momentum fund, every price gain feeding the next leg higher as more money joins the uptrend. Short positions held by industrial hedgers are stopped out, throwing further fuel on the flames.

Outside of China, only silver has seen similar mass surges in the past, most notably in 2021, when the Reddit message board launched a wave of retail buying in pursuit of a mythical market short.

No surprise then that silver has once again been the wildest of all markets in recent days.

But this phenomenon is spreading.

The late-January surge in the CME copper contract to a record high of $6.58 per lb wasn't accompanied by any outsize increase in managed money long positions.

Rather, the real action was taking place in the CME's smaller contracts aimed squarely at retail investors.

The micro copper, opens a new tab contract, which at 2,500 pounds is a tenth of the size of the main contract, saw volumes mushroom from 369,000 lots in December to 969,000 in January. That's equivalent to over one million metric tons of physical metal.

The micro gold, opens new tab contract has experienced an equally dramatic jump in activity after bursting into life around the middle of last year.

DELTA FORCE

The wave of investment money has been flowing into metal options as well as futures.

The CME's copper event option contract, which offers a simple binary punt on the underlying price, notched up volumes of almost 83,000 lots in December and January, more than the total traded since the product was launched in September 2022.

Options act as accelerators on already supercharged rallies.

With everyone looking to snap up call options, conferring the right to buy, sellers have to hedge their exposure by buying into a rising market.

Such delta-hedging creates a mechanistic feedback loop, which runs until the momentum turns, at which point the process starts working in reverse as those who sold the options sell back their cover in a falling market.

The whiplash can be extreme, as silver investors have just found out.

LIQUIDITY TRAP

Animal spirits, compounded by options leverage, created the conditions for both the wild upswing in precious metals prices and the subsequent violent unwind.

Gold should be cushioned against such speculative storms by central bank reserves, but it too is ultimately a finite physical commodity.

Analysts at Citi calculate that were investors to increase their purchases of gold from the current $300-400 billion per year to $2 trillion, the price could exceed $10,000 per ounce.

That may sound like a lot of money, but an increase from $1 trillion to $2 trillion would represent just one six-hundredth of global household wealth, according to Citi.

If even gold is vulnerable to the raw power of money, consider the potentially destructive impact on smaller industrial metal markets such as nickel or tin.

That's why China, the world's foremost industrial metals producer and consumer, has been taking ever more stringent measures to prevent paper market wildness from contaminating real-world supply chains.

THEME AND MEME COLLISION

January's perfect metallic storm marks the collision of two powerful investment themes.

The fear of dollar "debasement" is causing both institutional and individual investors to diversify into harder assets.

Metals, meanwhile, were already attracting investor interest due to their central role in both energy transition and Internet of Things mega-trends.

Silver, which is both a precious and industrial metal, has been the pivot between macro- and micro-investment narratives.

In the new world of social media, these colliding metallic themes have morphed into internet memes, amplified by message boards such as WeChat in China and Reddit in the West.

The result is unprecedented investment appetite for metals, both precious and base, at global scale.

And it's unlikely to be sated any time soon.

More turbulence seems guaranteed until the fever breaks.

If it breaks.

Bitcoin’s Weekly Sell-Off: Key Levels Doing the Talking

Bitcoin’s recent weekly sell-off has been eye-catching — it’s tracking toward one of the largest weekly declines we’ve seen since late 2022, a level of downside velocity not commonly seen outside major market corrections.

That said, the market isn’t moving aimlessly — structure still matters.

1️⃣ First support: ~65,000 (Weekly Anchored VWAP) This level — anchored from the 2022 bear market lows — has been a focal point for buyers during the pullback. A weekly close below this anchored VWAP would be technically notable, as it has defined trend support throughout the post-2022 recovery.

2️⃣ Deeper confluence: ~58,000 (200-week MA + 61.8% Fib) Beneath the anchored VWAP lies a key structural area where the 200-week moving average overlaps with the 61.8% Fibonacci Retracement of the 2022 lows to all-time highs range.

Confluences like this stand out — These are areas where markets often pause, reassess, and reveal longer-term intent…particularly during shifts in risk tolerance.

What to watch next:
- A weekly close above the VWAP keeps the long-term uptrend intact.
- A close below it signals potential rotation toward deeper support at the 200-week MA / Fib zone.
*Not a recommendation

, opens new tab