AMD JUST MADE ITS AI AMBITIONS HARDER TO IGNORE

Lisa Su used AMD's Advancing AI conference in San Francisco to show that the company is no longer simply trying to catch Nvidia. AMD is building a complete AI platform — chips, processors, software, networking and full data-center systems.

Here are the biggest announcements she discussed:

Helios is now in full production. AMD's new rack-scale AI system is being manufactured, with customer shipments expected to begin by the end of Q3. Helios combines AMD GPUs, EPYC processors, networking and software inside one complete system.

AMD wants a larger share of the entire AI data center. The company is no longer focused on selling customers only an individual chip. AMD wants to provide the complete infrastructure needed to build and operate large AI systems.

OpenAI plans to use AMD at a much larger scale. OpenAI expects to begin deploying Helios systems toward the end of 2026, with deployments increasing throughout 2027. It also expects to use AMD's future MI500 accelerators.

AMD now has support from several major AI customers. OpenAI, Microsoft, Meta and Anthropic are all working with AMD — valuable customer validation that the largest AI companies want alternatives to relying only on Nvidia.

The new MI455X performance claims were impressive. AMD said the accelerator could deliver significantly higher token throughput than its previous generation, and Helios could provide better performance for the money.

ROCm.AI is designed to improve AMD's software. A new software platform intended to make AMD chips easier to program, optimize and deploy — important because Nvidia's CUDA has traditionally been one of AMD's biggest competitive challenges.

AMD is partnering with Cerebras. Combining AMD's Helios systems with Cerebras' wafer-scale AI processors for faster AI inference and extremely large AI workloads.

The next-generation Venice processor is moving forward. Lisa Su discussed AMD's upcoming EPYC "Venice" server processor, which will become an important part of future AMD data-center and AI systems.

AMD plans to release a new rack-scale platform every year. Helios is not a one-time product. AMD is building a long-term AI roadmap with future processors, accelerators, software and complete systems already under development.

Why This Conference Mattered

The market may have wanted one enormous revenue number or a surprise announcement that would immediately send the stock higher. Instead, Lisa Su showed that AMD is executing, putting products into production, improving its software and attracting some of the largest AI customers in the world.

Not Subscribed Yet?

Sign up for the newsletter here

The AI Trading Institute

JULY COHORT: SOLD OUT

The July 25 cohort is full. Add your name to the August waitlist and we'll notify you the moment enrollment opens.

WHY TESLA IS GETTING HIT SO HARD

Revenue was strong, but Wall Street focused on weaker profits, heavy spending and the uncertain timing of Tesla's next growth engines.

The Quarter Was Not Bad on Revenue

Tesla reported $28.24 billion in revenue, up 26% from a year earlier and ahead of Wall Street's expectations. Vehicle deliveries were also strong. The problem was that the growth did not translate into the level of profit investors expected.

Tesla earned about $0.33 per share, compared with expectations near $0.53. Net income was approximately $1.1 billion. In plain English, Tesla sold more, but the additional sales did not produce enough additional profit to satisfy the market.

Tesla Is Spending Enormous Amounts of Money

The company generated cash from its operations, but it spent roughly $5.8 billion during the quarter on factories, equipment, AI, Cybercab, robotaxis and Optimus. That pushed free cash flow to approximately negative $1.1 billion — Tesla's first negative free-cash-flow quarter in more than two years.

Management now expects capital spending to exceed $25 billion in 2026. Tesla is not running out of money; it is investing heavily today in businesses it believes can become much larger tomorrow.

The Car Business Is Still Under Pressure

Tesla's automotive business remains its financial foundation, but profitability is being squeezed by lower selling prices, intense competition and weaker regulatory-credit revenue. Regulatory-credit revenue fell sharply to about $146 million, removing a source of high-margin income that has helped earnings in the past.

R&D spending climbed to approximately $2.37 billion as Tesla increased investment in AI, autonomy and robotics. Those investments may matter for the future, but for now they are adding costs faster than profits.

Cybercab Progress Was Real — But Not Enough

Tesla has started producing Cybercabs in Texas, expanded its robotaxi efforts and reported continued growth in Full Self-Driving subscriptions. Those are real signs of progress.

Wall Street wanted clearer answers about how quickly the service can expand, how much it will cost and when it will produce meaningful profits. Investors can see the spending today, but they still cannot clearly see the payoff. That uncertainty is a major reason the stock is being punished.

INTEL FINALLY SHOWS SIGNS OF A REAL TURNAROUND

A cleaner quarter, stronger AI demand and better guidance — but cash flow still matters.

Intel reported a much stronger quarter than Wall Street expected, beating on revenue, adjusted earnings, margins and guidance. Revenue rose 25% from last year, while adjusted earnings came in at twice the estimate.

Earnings Scorecard

AI Was the Standout

Data Center and AI revenue jumped 59% to $6.3 billion. Intel said server-chip demand was stronger than expected and exceeded available manufacturing capacity in some areas. The company has also signed several long-term customer agreements for data-center processors.

The PC Business Improved

PC and Physical AI revenue rose 13% to $8.9 billion. Unit shipments were lower, but stronger pricing and a greater mix of premium processors lifted revenue.

Guidance Stayed Strong

Intel expects Q3 revenue of $15.8B to $16.8B. It also projected adjusted EPS of $0.38 and an adjusted gross margin of 42%.

What About the $11 Billion Loss?

The GAAP loss was mainly caused by a $12.53 billion accounting charge tied to escrowed shares and the government investment agreement. Excluding special items, Intel earned $2.2 billion, or $0.42 per share. The core business did not lose $11 billion.

Market Tidbit

WALL STREET BETS $175 MILLION ON SALAD OIL

A market story so strange it sounds impossible.

Wall Street has been fooled by fake earnings, bad accounting and brilliant con artists. But in 1963, some of the biggest financial institutions in America were fooled by something much simpler: oil floating on top of water.

Commodity trader Anthony "Tino" De Angelis claimed to control enormous tanks filled with valuable soybean oil. Banks and lenders accepted warehouse receipts showing that the oil existed and loaned him millions of dollars against it.

The paperwork looked legitimate. The tanks looked full. The oil was supposedly sitting right there. Except much of it was not oil at all.

Many of the tanks were reportedly filled mostly with water, with only a thin layer of oil floating across the top. Inspectors lowered their testing equipment, hit oil first and assumed the entire tank was full.

De Angelis allegedly made the deception even more convincing by pumping the same oil from one tank to another during inspections, allowing the same inventory to be counted again and again.

The Bet Gets Bigger

With millions of dollars borrowed against inventory that barely existed, De Angelis began making massive bets in the soybean oil futures market. For a while, it worked. Then inspectors finally discovered the truth. Roughly $175 million worth of supposed inventory had vanished because much of it had never existed in the first place.

American Express had guaranteed many of those warehouse receipts through a subsidiary. When the fraud collapsed, the company faced enormous losses and its stock was crushed. Investors ran for the exits.

Warren Buffett saw something different. He reportedly checked whether customers had stopped using American Express products. They had not.

Buffett concluded that Wall Street was confusing a terrible scandal with a destroyed business. He bought a major position in American Express while nearly everyone else was panicking — and the investment became one of the most famous wins of his career.

Sometimes the craziest market stories are not about technology or complexity. They are about greed, bad assumptions and panic.

Week of July 20 — Two /ES 0DTE trades

Two /ES 0DTE plays — morning and late-day. Same strategy family, different entry times.

Free Community

Join 850+ traders in our Discord

Trade ideas, screenshots, daily banter. Zero cost.