Trading 212 £190k portfolio update 2026-05-29

 A bumper week, mainly due to positive news on the Hormuz Strait.

My 1 yr IRRs are 47% (GIA) and 43% (ISA)
My all time IRRs (approx 2 yrs) are 95% (GIA) and 55% (ISA)

I decided to spend a little of the spare cash I held in the accounts to top up on the following:

ServiceNow seemed to be trending (several YT channels were tipping it), so I investigated and then bought £500 inside my ISA and then watched it go up 20% and then set a stop loss to make a quick profit should it start to drop. I review my stop loss levels every few days and adjust them upwards when necessary.

I think the USA economy is headed for hard times and my gold/silver miners seem cheap at the moment, so I have topped up on some gold miners. Most people expect gold to go to $5000 or even over $6000 by the end of 2026. At worst this would be a 5% rise, at best 15% - not as good as tech stocks, but it may act as a small hedge.

I am interested in the smaller miners. Due to Hormuz, running costs have dramatically increased and so production may have been deliberately cut back and expansion delayed while energy is expensive.

The "Lucrative Spread" Profit Boom: Because running costs have stabilized near $1,600/oz while the market price of gold has soared past $4,500/oz, gold mining companies are currently experiencing the widest profit margins in their history. Operators are generating massive free cash flow, allowing them to rapidly pay down debt, fund new projects, and increase dividend payouts to stock investors.

Junior miners may be experiencing cash flow problems and cash-rich larger miners may buy them up. This would increase their efficiency, reduce costs and increase profits. 

At least that is my theory!

This is not investment advice. My T212 portfolio is not my only portfolio and it is my 'higher risk, higher reward' portfolio, not a sensible, slow growth 'portfolio for life'.

I try to keep the stocks that I intend to keep for a long time in my GIA Invest account, and the more volatile (risky) stocks in my ISA so I can sell without triggering a taxable gain.

If you are new to investing, you could look at some of my other blog articles (e.g. simple two-ETF portfolios) for more sensible ideas.

Setting stop losses

I have set a number of stop losses on positions inside my ISA account. If I am up by 15% on my more volatile stocks, I will usually set a stop loss at -10% (or -15% if the stock is very volatile). I am fully expecting a big dip. I revise the stop loss limits regularly.

Setting a stop loss on my GIA account stocks is not quite so desirable because of the tax loss it would incur. For this reason, if I do set a stop loss within the GIA, I either set it at around the break-even point or I set it to sell just a small amount to trigger an event. I will then examine that stock manually to see if I should sell more/all of it.

In the past, I have had bad experiences with stop losses on T212. They have been triggered by extremely short-lived market fluctuations - almost is if it has been done deliberately. Setting a stop loss tells the whole world (or just the main broker which is Interactive Brokers for T212) that they can pick up your shares cheaply. All they have to do is sell a load of the same shares which will bring down the market price, this will then hit your stop level, you will be sold out and they can buy all their shares and yours back again at a cheaper price. If I was a suspicious person, it's as if Interactive Brokers just need to process all their sell orders on their books first, trigger all the stop losses on their books, and then process their buy orders to pick up the shares for cheap (having already transacted them at the previous 'open book' rate for their buy customers).

Trends and Tips

The best performers so far this year has been tech stocks (disks, memory, chips, S. Korea). However, it seems the biggest factor now for AI data centers is the lack of energy (electricity). There is a shortage of generators and infrastructure. $Billions are tied up in hardware, data centers are being built, but they cannot connect them to juice!

AI running costs are rising and AI companies are ramping up their prices. Businesses have got rid of  10%+ of their workforce and replaced them with AI subscriptions, but now they are finding it very expensive. Their AI provider can simply raise their prices and they have the company over a barrel.

Spike costs are triggering a massive enterprise migration toward hybrid and on-premises infrastructure. Instead of renting massive cloud models, companies are buying standardized, out-of-the-box corporate servers from companies like Dell or Lenovo. They use this hardware to host powerful, free open-source models (like Meta's Llama series) internally. This gives them predictable, flat-rate infrastructure costs and total data sovereignty. 

This may account for the massive rise in Dell and Lenovo recently.

• Insatiable Order Backlogs: This is not speculative future demand; it is hard cash on the table. Dell has an ironclad $51.3 billion AI server backlog, while Lenovo is sitting on a $21 billion sales pipeline. Their production lines are completely booked out into next year, ensuring highly predictable near-term revenue growth
• The PC Cycle Safety Net: Even if AI server demand cools slightly, both companies are on the cusp of a major corporate commercial PC refresh cycle. Millions of enterprise laptops bought during the 2020/2021 remote-work boom are hitting retirement age, guaranteeing a strong operational safety net for their legacy divisions.

This trend will continue. Demand will be from large businesses initially who are finding their AI subscription costs are too expensive and PC/laptop refresh (Win 11).

I don't see this trend slowing down. Dell and Lenovo may be too expensive at the moment but if Dell comes down from 420 to 400, I may have a nibble. Even at the current prices however, they could rise 25% by next year.

Beyond Dell and Lenovo, several key hardware manufacturers build AI-optimized servers, competing aggressively for enterprise and cloud provider contracts:

• Super Micro Computer (Supermicro / SMCI): Dell's most ferocious competitor. Supermicro specializes in ultra-fast time-to-market hardware, often shipping systems with Nvidia’s latest chips weeks before competitors. They are highly favored for their modular "building block" architecture and advanced liquid-cooling solutions designed for massive clusters.
• Hewlett Packard Enterprise (HPE): HPE targets massive enterprises and sovereign governments with its Cray Supercomputing line and ProLiant AI servers. They specialize in massive, highly secure deployments and offer a unique hardware-as-a-service consumption model called HPE GreenLake.
• Cisco Systems: Known historically for data center networking, Cisco builds the UCS (Unified Computing System) servers. They specialize in integrating Nvidia computing power directly with enterprise networking fabrics and AI-ready security systems.
• Foxconn (Hon Hai Precision Industry): While consumers know Foxconn for assembling iPhones, they are actually the underlying manufacturing backbone for a massive percentage of the world's AI servers. They build custom, white-label server motherboards and infrastructure directly for big tech hyperscalers.
• Inspur: The dominant player in the Asian market. Based in China, Inspur provides high-density AI servers and command structures tailored for major enterprise workloads across Asia-Pacific.

SMCI, HPE and Cisco have also spiked in the last week. Hon Hai Precision not as much (+29% in 1 month). All are available on T212. SMCI try to undercut Dell and so have small margins and heavy debt. They are a risky bet (but maybe trendy and so liked by the retail share buyers!).

These companies, together with Dell and Lenovo may be worth watching. HPE and especially Cisco (my best choices as I like picks-and-shovel companies!) seem undervalued even at the current price and both are on my radar...

---