HOME > BLOG

The Best and Worst State Government Websites in America

Nemanja
Nemanja
February 25, 2026
  • Tennessee's mobile site takes 24 seconds to load, the worst in the country by a wide margin.
  • Iowa is the only state to score perfectly across every quality category on both desktop and mobile.
  • Oklahoma loads in 1.28 seconds on both platforms but ranks 41st overall.

State government websites are among the most visited public-sector sites in the country. They handle everything from business licensing to public health information, and for many residents they are the primary interface with state government.

Despite that, web performance across state sites varies enormously. Some states have invested in modern infrastructure that loads quickly and scores well across every technical benchmark. Others have not.

We tested all 50 state government homepages plus Washington D.C. using Google PageSpeed Insights, measuring speed and overall quality across desktop and mobile. The results show a wide gap between the best and worst performers, with no clear regional or population pattern explaining who gets it right.

The Fastest State Government Websites

Oklahoma is the only site that loads in exactly the same time on desktop and mobile: 1.28 seconds each. Michigan has the fastest raw desktop score at 1.23 seconds with a solid 2.68 seconds on mobile. Georgia is close behind at 1.38s desktop and 2.44s mobile, the second-fastest mobile score in the country.

Iowa and Kentucky stand out for combining speed with strong overall scores. Iowa loads in 1.27s on desktop and 3.3s on mobile and carries a total average score of 96.96. Kentucky's desktop performance score of 99 is the highest in the country.

California has the single fastest desktop score in the dataset at 1.05 seconds but a more middling mobile score of 5.46s, placing it outside the combined top 10.

RankStateDesktop SpeedMobile SpeedCombined Avg
1Oklahoma1.28s1.28s1.28s
2Georgia1.38s2.44s1.91s
3Michigan1.23s2.68s1.96s
4South Dakota1.31s2.90s2.10s
5Kentucky1.33s3.06s2.20s
6Iowa1.27s3.30s2.29s
7New York1.40s3.70s2.55s
8Maine1.71s4.23s2.97s
9Arizona1.70s4.50s3.10s
10New Hampshire1.70s4.60s3.15s

The Slowest State Government Websites

Tennessee is the worst performer by a wide margin. Its desktop Speed Index of 8.53 seconds is the highest in the country, and its mobile score of 24.27 seconds is nearly double the next-worst state. Virginia and South Carolina both exceed 6 seconds on desktop and 16 seconds on mobile. Kansas, Indiana, and New Jersey round out a bottom tier where mobile load times run between 14 and 17 seconds.

RankStateDesktop SpeedMobile SpeedCombined Avg
Tennessee8.53s24.27s16.40s
2Virginia6.78s16.74s11.76s
3South Carolina6.09s16.96s11.53s
4Kansas5.87s16.68s11.28s
5Indiana5.61s16.25s10.93s
6New Jersey5.53s14.07s9.80s
7Delaware4.93s14.01s9.47s
8Montana4.87s13.84s9.36s
9Hawaii4.47s13.05s8.76s
10North Dakota2.75s13.06s7.91s

Overall Performance: All States Ranked

Total score is the average across all eight metrics: desktop and mobile performance, accessibility, best practices, and SEO. Iowa leads by a significant margin, scoring perfectly across accessibility, best practices, and SEO on both platforms. South Carolina ranks last, with the lowest accessibility scores in the dataset. Colorado scores 35 on best practices on both platforms, the lowest in that category, while still hitting 100 on SEO. Tennessee, the slowest site in the country on speed, manages a perfect 100 on desktop accessibility, which illustrates how unevenly these scores can be distributed across categories.

RankStateTotalDesk PerfMob PerfDesk AccessMob AccessDesk BPMob BPDesk SEOMob SEO
1Iowa96.969498100100100100100100
2Rhode Island91.38797598981001009292
3Arizona90.759574100100100100100100
4Maine89.6797921001007373100100
5Pennsylvania89.637268100100100100100100
6Arkansas89.0463591001009696100100
7Michigan88.67957410096100100100100
8Massachusetts88835999921001009292
9Vermont87.1762569596100100100100
10Maryland87.04695685829292100100
11Connecticut86.928061969696968585
12Louisiana86.88855098989696100100
13New York86.7184687773100969292
14Georgia86.67978979801009692100
15California86.178668828296967373
16Kentucky85.799998737392928080
17Oregon85.63626091919692100100
18Hawaii85.4640359797100100100100
19Nebraska84.178482727877927785
20Missouri82.967248100100100969292
21Florida82.254960949377778585
22Utah81.67854210010096929292
23Minnesota81.4666439789100969292
24Illinois81.387057908696929292
25West Virginia81.176963979792968585
26New Jersey81.136743969096969292
27Virginia812837929596968585
28New Hampshire80.7981561001005454100100
29South Dakota80.7985429895100100100100
30Alaska80.56848979777779292
31North Carolina80.337845838292889292
32New Mexico80.1763319693100100100100
33Wisconsin79.63835293938177100100
34Washington79.04563110092969610092
35Ohio78.59447959892968585
36Idaho78684778811001007777
37Delaware77.8863359710096927373
38Mississippi77.08835796966262100100
39Nevada76.887557929192927777
40Montana76.045555999473737777
41Oklahoma75.63534510010073738585
42North Dakota75.466544838273738585
43Alabama74.83423596965454100100
44Kansas74.582841939373778383
45Wyoming74.085054857577738383
46Texas746253848850466977
47Colorado73.29464096913535100100
48Indiana73.135643959177779292
49Tennessee73.0851431008773738282
50District of Columbia71.889349879177778282
51South Carolina71.796753737973738383

Methodology

All scores were collected using Google PageSpeed Insights (pagespeed.web.dev), which runs each URL through Google's Lighthouse auditing engine and returns scores across four categories: performance, accessibility, best practices, and SEO. Each category is scored from 0 to 100.

Performance measures how quickly a page loads and becomes usable, including metrics like First Contentful Paint, Largest Contentful Paint, and Speed Index. Speed Index specifically measures how quickly visible content populates on screen, scored in seconds. Lower is better. Accessibility tests whether a site can be used by people with disabilities, covering things like screen reader compatibility, color contrast, and keyboard navigation. Best practices checks for general web hygiene: HTTPS usage, correct image sizing, absence of known security vulnerabilities, and similar technical standards. SEO evaluates whether the page is structured in a way that search engines can correctly read and index.

Each site was tested separately on desktop and mobile, which use different simulated hardware and network conditions. Mobile tests simulate a slower device and network connection, which is why mobile scores are consistently lower than desktop scores across all states. The total average score is the mean of all eight individual scores for each state.

The homepage of each state's official government portal was tested. Results reflect a single point-in-time measurement and can vary depending on server load, content delivery network performance, and changes to the site after the date of testing.

PC Gamers Accidentally Became Better Investors Than Wall Street

A $187 computer component returned 123% in two months. The stock market returned 1.86%.

If you'd told investors in October that the smartest play for Q4 wasn't the S&P 500, wasn't gold, and wasn't semiconductors, but a 32GB kit of desktop RAM, they'd have laughed you out of the room.
They're not laughing now.
Between October 21 and December 24, 2025, a Crucial Pro DDR5 RAM kit went from $186.99 to $417.95. That's a 123.5% return in two months.
The S&P 500 returned 1.86%.
That means RAM outperformed the benchmark index by a factor of 66.Here's the kicker: RAM is a semiconductor. And it outperformed every major semiconductor ETF by at least 37x. Buying the actual chip beat buying the chipmakers.

ram price up

Why RAM? Why now?

The culprit is AI.
Samsung, SK Hynix, and Micron control over 90% of global DRAM production. For the past year, these manufacturers have deliberately prioritized HBM (high-bandwidth memory) production for AI data centers over consumer DDR5. The math is brutal: HBM consumes roughly three times the wafer capacity of DDR5 per gigabyte. Every AI chip made directly reduces consumer RAM supply.
Micron went furthest: it announced plans to shut down its Crucial consumer brand entirely by early 2026, with executives explicitly stating the move would "free up wafer supply for strategic accounts."
The result: a genuine shortage. Retailers in Japan are rationing RAM to one kit per customer. Framework hiked memory upgrade prices by 50%. A 64GB DDR5 kit now costs more than a PlayStation 5. A leaked SK Hynix internal analysis forecasts PC DRAM supply trailing demand until at least late 2028.
Some are calling it "RAMageddon."

Asset

Dollar Value on October 21, 2025

Dollar Value on December 24, 2025

Profit

SPY (S&P 500 ETF)

671.44

683.92

1.86%

Bitcoin (BTC-USD)

110,587.63

87,431.09

-20.94%

Brent Crude Oil

60.91

61.99

1.77%

Gold (XAU/USD)

4,332.70

4,516.70

4.25%

Crucial Pro DDR5 RAM 32GB Kit

186.99

417.95

123.51%

Soxx

292.5

302.21

3.32%

VanEck Semiconductor ETF (SMH)

347.06

358.6

3.33%

State Street SPDR S&P Semiconductor ETF (XSD)

348.72

329.72

-5.45%

Invesco QQQ Trust

611.64

618.2

1.07%

iShares 20+ Year Treasury Bond ETF (TLT)

91.99

87.07

-5.35%

iShares Silver Trust (SLV)

45.19

63.68

40.9

We've seen this before

If this sounds familiar, it should. From 2020 to 2021, crypto miners caused a similar crisis in the graphics card market. RTX 3080s that launched at $699 were reselling for $2,000 to $3,500. Bots scraped inventory within seconds. Miners spent an estimated $15 billion hoarding GPUs.
But there's a crucial difference. During the GPU shortage, miners were competing with gamers for the same product. The cards existed. They were just being bought by someone else.
This time, the product is being deliberately deprioritized. Manufacturers chose AI over consumers because that's where the money was. You can't outbid a bot if nothing's on the shelf, and nothing's on the shelf because the factory is making something else.

The accidental asset class

Nobody buys RAM as an investment. But anyone who happened to purchase DDR5 in early Q4 is now sitting on returns that would make most hedge funds jealous.
It's a strange time when computer components outperform financial instruments, but maybe it's not that strange. In an economy increasingly defined by AI infrastructure, the bottlenecks are the value.
The real alpha was in the PC parts aisle all along.

Methodology

We compared prices on October 21 and December 24, 2025. Financial assets were tracked using Yahoo Finance. RAM pricing is based on the Crucial Pro DDR5 32GB Kit (2x16GB, 6400MHz) on Amazon.

This analysis was compiled by Hirocom, a marketing agency that occasionally finds investment advice in unexpected places. This is not financial advice. Please do not liquidate your 401(k) to buy RAM.

Sources
TrendForce, "64GB DDR5 RAM Reportedly Now Pricier Than a PlayStation 5 Amid Soaring Memory Costs" (November 2025)
Tom's Hardware, "The RAM pricing crisis has only just started, Team Group GM warns" (December 2025)
Tom's Hardware, "Here's why HBM is coming for your PC's RAM" (December 2025)
TechRadar, "Why is RAM so expensive right now?" (December 2025)
Tom's Guide, "RAM price crisis LIVE" (December 2025)
TweakTown, "Samsung shifts focus from HBM to DDR5 modules" (December 2025)
Resell Calendar, "The RAM Price Crisis is Getting Worse" (December 2025)
Counterpoint Research, DRAM market share data (October 2025)
Jon Peddie Research, GPU shortage and crypto mining analysis (2021-2022)
Yahoo Finance

 

Blog Post
Subscribe to Blog Post