Warren Buffett is sitting on $325 billion in cash – his largest hoard ever.

Not because he wants to – but because he can’t find value in the usual places.

Now, as US government spending spirals out of control, Buffett knows he’s losing billions of dollars to inflation. 

That’s why I predict Buffett’s next investment will catch millions of people off guard. 

It’s not another bank… railroad company… or more shares of Apple. 

It’s a gold company. How do I know?

Because the math doesn’t lie:

You can buy the average gold developer for $30 and get back $13 a year —

That’s a 43% ROI annually.

Over 10 years, that’s $130 on a $30 investment.

Tell me where else Buffett can get that.

But there’s one specific miner Buffett likes best:

• It’s the best-managed major gold miner in the industry…
• Has massive cash flow…
• Is trading at a deep discount to fair value…
• Positioned at the heart of Trump’s new mining push…

Don’t wait for Buffett to reveal his position in his 13F filing on February 17th…

Right now, you have the chance to front-run the greatest investor of all time. Go here and I’ll give you the name and ticker – along with details on my top four small miners.

To your wealth,

Garrett Goggin, CFA, CMT
Chief Analyst & Founder, Golden Portfolio

P.S. A lot of investors write in to tell me how much they’ve made in Bitcoin. My reply? Good for you. First off, gold investing is cyclical. You really only want to own gold at one specific time in the cycle. That time is now. Second, the world’s governments are not buying Bitcoin. They’re betting on gold. All of them. Bitcoin (does anyone really know for sure the US government didn’t create it?) will be a good bet… until it isn’t. It may end up doing great. Or it may be eclipsed by any number of tech developments. 

Meanwhile, gold will continue to do what it’s done for almost 6,000 years of recorded human history: Protect wealth through chaos. Go here if you want the name and ticker of Buffett’s likely gold play… and details on my top four miners

---

Just For You

Microsoft’s Maia 200: The Profit Engine AI Needs

Written by Jeffrey Neal Johnson. Date Posted: 1/27/2026. 

At a Glance

• Microsoft’s new custom silicon chip is designed to significantly reduce the cost of running artificial intelligence workloads for the cloud infrastructure division.
• Management timed this strategic hardware release to reassure investors about profit margins just before the fiscal second-quarter earnings announcement.
• Moving inference processing to proprietary hardware allows the tech giant to depend less on third-party suppliers and to improve long-term cloud economics.

Microsoft (NASDAQ: MSFT) officially launched its custom Maia 200 AI accelerator in the last week of January, marking a milestone in the company’s infrastructure strategy. The announcement arrives at a critical moment for the tech-sectorgiant—just 48 hours before management is scheduled to release its fiscal second-quarter earnings report.

For investors, the timing is a deliberate signal. Over the past year, Wall Street has taken a “show me” stance on Microsoft, which is trading near $470. Although shares have recovered from recent volatility, concerns persist about the massive capital spending required to build out AI data centers.

The biggest scam in the history of gold markets is unwinding (Ad)

There are 90 paper gold claims for every real ounce in COMEX vaults. Ninety promises, one ounce of metal. It’s like musical chairs with 90 players and one chair. COMEX gold inventory dropped 25 percent last year alone as gold flows East to Shanghai, Mumbai, and Moscow. On March 31st, contract holders can demand delivery. When similar situations arose in the past, markets closed and rules changed. Paper holders got crushed while mining stock holders made fortunes. One stock sits at the center of this crisis.Get the full story on this opportunity now.

By unveiling a proprietary chip optimized for inference immediately before reporting results, management is signaling a shift: the focus is moving from expanding AI capacity at any cost to optimizing it for long-term profitability.

3nm Power & Speed: Why Specs Matter

To gauge the financial impact, investors should start with the technology. The Maia 200 is built on Taiwan Semiconductor Manufacturing Company’s (NYSE: TSM) advanced 3-nanometer process, packing more than 140 billion transistors onto a single chip and pairing that with 216GB of high-bandwidth memory (HBM3e) for rapid data throughput.

More important for shareholders than transistor count is purpose: the Maia 200 is specifically optimized for inference workloads.

The Difference Between Learning and Doing

AI has two main phases:

• Training: Teaching an AI model, which requires enormous computation and is typically done with general-purpose GPUs such as those from NVIDIA (NASDAQ: NVDA).
• Inference: The AI’s day-to-day operation. Every time a user asks Copilot a question or uses ChatGPT, the system performs inference to generate an answer.

Training is a massive upfront cost; inference is a recurring and perpetual expense. As millions adopt Microsoft’s AI tools, inference becomes a dominant, ongoing cost. Deploying a chip tailored to that task allows Microsoft to handle daily interactions faster and more cheaply than with third-party hardware.

Economics of AI: Turning Efficiency Into Profit

The headline from the announcement is that the Maia 200 delivers roughly 30% better performance per dollar versus Microsoft’s prior hardware configurations. For CFOs and institutional investors, that’s the most consequential figure.

This improvement directly affects Cost of Goods Sold (COGS) for Microsoft’s cloud business. In software, gross margins are a primary measure of financial health. If Microsoft relied entirely on expensive third-party hardware to run its services, growing usage would compress margins. Cutting the cost of each AI query by about 30% with its own chips can materially expand gross margins on subscription services like Microsoft 365 Copilot and Azure OpenAI Services.

The Hidden Cost: Energy and Power

There’s a secondary benefit: lower electricity consumption. AI data centers are power-hungry, and a move to a smaller 3-nanometer architecture means the Maia 200 uses less energy for the same work as older chips.

Given Microsoft’s recent large energy commitments to power its data centers, reducing watts per query is nearly as important as reducing dollars per chip. That dual efficiency helps insulate the company from volatile energy prices and supports the bottom line.

Microsoft vs. The Field: Catching the Hyperscalers

The Maia 200 also changes the competitive picture among hyperscalers—Amazon Web Services (AWS) and Google Cloud Platform (GCP) among them. Both Amazon (NASDAQ: AMZN)and Alphabet (NASDAQ: GOOGL) have developed custom chips for years, which gave them a theoretical cost edge.

Today’s data suggests Microsoft has narrowed that gap. The company reports the new chip delivers:

• Three times the performance of Amazon’s third-generation Trainium chip on certain FP4 benchmarks.
• Superior performance versus Google’s seventh-generation TPU on FP8 precision tasks.

Achieving technical parity or superiority in custom silicon reduces the risk of losing price-sensitive enterprise customers to rivals.

Supply Chain Leverage

This move also gives Microsoft greater leverage. The industry has been constrained by NVIDIA GPU supply, and shortages and high prices have slowed growth for many customers.

While Microsoft will continue partnering with NVIDIA for AI training, the Maia 200 insulates the company from hardware bottlenecks for inference workloads. That helps Microsoft scale Copilot and other services without being limited by third-party hardware availability.

Custom Silicon & the Road to $600

The Maia 200 aligns with the bullish narrative on Wall Street. Analysts remain optimistic about Microsoft’s long-term outlook despite recent consolidation.

Firms such as Wedbush have described Microsoft as a front-runner in the Fourth Industrial Revolution and continue to maintain aggressive price targets above $600. The consensus among 30+ analysts is a Buy, with an average price target implying more than 30% upside from current levels.

The Maia 200 addresses a key bear case—that AI spending would permanently erode profits. By demonstrating cost reductions, Microsoft gives analysts more support for high price targets.

Investor Outlook: All Eyes on Earnings

Attention turns to Wednesday, Jan. 28, when Microsoft reports Q2 earnings. Consensus projects revenue above $80.28 billion, but the stock’s reaction will likely hinge on forward-looking guidance rather than past results.

Today’s announcement creates a favorable backdrop for that call. Management can now point to the Maia 200 as a tangible driver of improved AI yield and cost control.

The Maia 200 marks a transition: Microsoft is shifting from build-at-any-cost expansion to operational efficiency. For shareholders, that’s a bullish development. It suggests management has a clearer path to protecting margins as AI adoption scales. If the upcoming earnings report confirms strong demand for Azure and Copilot, the improved economics from the Maia 200 could help Microsoft retest prior highs and move toward the analyst-projected $600 target over time.

Thank you for subscribing to Earnings360, a morning newsletter that summarizes quarterly earnings for public companies that trade on U.S. markets.

This email message is a paid sponsorship provided by Golden Portfolio, a third-party advertiser of Earnings360 and MarketBeat. 

If you have questions about your newsletter, please feel free to email MarketBeat’s U.S. based support team at contact@marketbeat.com.

If you no longer wish to receive email from Earnings360, you can unsubscribe.

Copyright 2006-2026 MarketBeat Media, LLC. All rights reserved.
345 N Reid Pl., Sixth Floor, Sioux Falls, SD 57103. United States..

From Our Partners: Momentum Trackers Just Lit Up — Here’s Why (Click to Opt-In)